Scottish Government

CHAPTER 2 SOCIAL SECTOR RENTS AND SERVICE CHARGES IN SCOTLAND

2.1 Introduction

This chapter sets out the profile of council and housing association rents by size and local authority area for 2005/06. It also discusses the rent setting methods used by local authorities, and presents the more limited information available in respect of council and housing association service charges that are levied on tenants in addition to their rents.

With this comprehensive data set of council and housing association rents, it has been possible to generate disaggregated data for Scotland as a whole and to undertake more detailed analyses of the profile of average rents in the social sector in relation to size, location, house prices and earnings. This chapter first outlines each of the data sets and the key characteristics of the data. Second, it analyses local authority rent setting systems, the affordability of existing social sector rent levels and the relationship between local rent levels and the costs of house purchase.

2.2 Local Authority Rent Levels

Data on local authority rents by size of dwelling were drawn initially from the data submitted by Scottish authorities to CIPFA. ⁴ However, not all Scottish authorities submit returns to CIPFA, and in those cases rent data was obtained directly from the authorities themselves. These authorities were Aberdeenshire, East Dunbartonshire, Highland, Inverclyde, North Lanarkshire and West Dunbartonshire.

It was also apparent from a cross-check with aggregate rent data submitted by authorities to the Scottish Government that there were a number of anomalies and inconsistencies between the two sets of rent data. Further enquiries were made of the authorities where there was a significant anomaly, and clarification was sought concerning the sources and definitions for the data each authority had submitted to CIPFA (or the research team) and the Scottish Government respectively.

A serious anomaly was defined as a difference between the aggregate rent figures of more than £1 per week, and this affected 8 authorities. Some of the anomalies were easy to resolve. In 4 cases (Dundee City, Eilean Siar, Inverclyde and Stirling), the weekly rent figures submitted to CIPFA had not been converted to 52 week equivalents by authorities that, in practice, collected their rents over 48 weeks.

Other anomalies were more problematic and reflected differences in definitions relating to the available rents disaggregated by size of dwelling, as opposed to the overall rents. This applied in some cases, for example, in respect of service charges. The underlying problem is that the rent systems of some authorities are primarily operationally based and are not geared to providing analytical data. In that context, there is an unavoidable element of estimation in the rent data those authorities submit to CIPFA and/or the Scottish Government, and there are some variation to the sources and methods for that estimation.

For Angus, the higher CIPFA rent figures have been preferred because they are more consistent with the wider housing revenue account information submitted to the Scottish Government. The East Renfrewshire figures submitted to CIPFA have been scaled down to the aggregate level submitted to the Scottish Government on the basis that the CIPFA figures were estimates that, in some cases, included elements of service charges. For Edinburgh, the figures submitted to the Scottish Government were an uprated estimate, while the CIPFA figures were the product of a new report from their rent system. The latter has, therefore, been preferred. West Lothian provided revised figures for the rents of dwellings by size to replace those submitted to CIPFA, as the latter had included service charges and charges for furniture.

After making these adjustments, the rent figures for local authorities at April 2005, by size of dwelling, are set out in Appendix C, Table 1. This shows that there are very substantial differences in the rents charged by individual authorities, with the highest rents almost double the level of the lowest. This is broadly the case for each size of dwelling, as well as for the overall stock of each council. There are, however, even greater differentials in the rents of bedsit dwellings.

In overall terms, rents are lowest in Moray, although in the case of three-bedroom dwellings rents are lowest in Midlothian; and in the case of four(+)-bedroom dwellings, rents are lowest in East Lothian. More generally, the overall rents for these three authorities were some way below the rents for all other authorities, and they were the only authorities where overall average rents were less than £40 per week. The different rankings of the rents of the three authorities for different sizes of dwellings reflects differences in the size-related differentials used in each authority.

Similarly, in overall terms, rents are highest in Inverclyde, although three-bedroom rents are highest in Edinburgh and four(+)-bedroom rents are highest in the Shetland Islands.

The differentials in rent levels by size of dwelling are shown in Appendix C, Table 2. The average differentials are relatively limited, with average bedsit rents at around 80% of the rents for two-bedroom dwellings. Similarly, rents for one-, three- and four(+)-bedroom dwellings are some 92%, 108% and 119% of the rents of two-bedroom dwellings, respectively.

There are, however, very marked differences in the size-related differentials within the rent structures of individual councils. Thus, for example, there is very little difference in the rents for dwellings of different sizes in Clackmannanshire, East Ayrshire, North Ayrshire, Perth and Kinross, and Stirling. In each case, rents for four(+)-bedroom dwellings are less than 10% higher than the rents for two-bedroom dwellings; and apart from North Ayrshire (which has no bedsits) and Stirling, rents for bedsits are less than 10% lower than the rents for two-bedroom dwellings. Within this group, the flattest rent structure by far is in North Ayrshire, where an additional room adds less than 1% to the average rent.

At the other extreme, rents in the Shetland Islands for four(+)-bedroom dwellings are some 70% higher than those for two-bedroom dwellings. Only in five other authorities were rents for four(+)-bedroom dwellings more than 30% higher than those for two-bedroom dwellings (Aberdeenshire, Angus, Dundee City, Moray and Renfrewshire). Four other authorities had rents for four(+)-bedroom dwellings between 20% and 30% higher than those for two-bedroom dwellings, while 11 authorities had rents between 0% and 20% higher than those for two-bedroom dwellings.

Figure 2.1 shows the differences in average rents for two-bedroom dwellings, which range from £35.51 per week in Moray up to £55.51 per week in Inverclyde.

Figure 2.1: Council rents at April 2005 (for two-bedroom dwellings)

2.3 Housing Association Rent Levels

Data on 2005/06 housing association rent levels have been taken from the Communities Scotland website, and are set out in Appendix C, Table 3. The figures shown are for secure tenancies only, and do not include the very small number of lettings subject to 'fair rents' set by the rent service. Appendix C, Table 3 also shows, separately, the figures for the 3 LSVT landlords (Dumfries and Galloway, Glasgow, Scottish Borders), because otherwise they would dominate the figures for other housing associations operating in those 3 areas.

The national average housing associaion rent at April 2005 is £50.27 per week. This compares with an average rent figure for council housing in Scotland in that year of £45.02 per week. However, it should be recognised that there are significant differences between the sectors in the distribution of the stock in terms of size, condition and location. There are also differences in accounting practices in terms of the separation of service charges and rents. These factors mean that this simple 'headline' comparison needs to be treated with some caution.

What is very clear, however, is that there is much less variation in housing association rent levels across Scotland than is the case with council rents. This is even clearer if the relatively low rents of the LSVT landlords in Dumfries and Galloway and Scottish Borders are set to one side. For two-, three- and four-apartment dwellings, the rents in the highest local authority areas are less than a quarter more than in the lowest areas. For one- and five(+)- apartment dwellings, the range is a little more; but even in those cases, the rents in the highest areas are no more than a half higher than those in the lowest areas.

Figure 2.2 shows the differences in average housing association sector secure rents for three-apartment dwellings across Scotland. These range from £46.99 per week in West Dunbartonshire to £57.76 per week in Inverclyde. This clearly shows far smaller variation in rent levels by area, compared to the local authority rents (shown in Figure 2.1).

Figure 2 .2: Housing association secure rents 2005/06 (for three-apartment dwellings)

However, there is a further level of variation in rent levels within the housing association sector - the differences between the rents of each association operating in each area. Figure 2.3 shows the range of the average secure rents for three-apartment dwellings for each association within each local authority area.

Figure 2.3: Variations in local housing association secure rents (for three-apartment dwellings, 2005/06)

There are some very large variations in average rent levels between individual associations in some local authority areas. These very large variations are entirely attributable to figures for individual associations and are sharply differentiated from the rest of the sector. Moreover, in all such cases, the number of dwellings involved in each area is very small (i.e. less than 10 dwellings). For the great majority of housing association stock in each area, the variations in the average rent levels for each association is far less pronounced.

The size-related differentials in housing association rent levels are shown in Appendix C, Table 4. The differentials are, on average, rather greater than in the council sector, with rents for one-apartment dwellings some 78% of the rents for three-apartment dwellings. Rents for two-, four- and five(+)-apartment dwellings are, respectively, some 88%, 113% and 127% of the rents for three-apartment dwellings.

While there are some fairly substantial local variations in the size differentials, they are far less extreme than those found in the council sector. Only in 3 areas are rents for five(+)-apartment dwellings more than double those for one-apartment dwellings, and only in 4 areas (as well as for one LSVT landlord) are rents for five(+)-apartment dwellings less than a half higher than those for one-apartment dwellings.

2.4 Local authority rent-setting systems

The study has also updated the information from the 2003 Glasgow study in respect of the rent setting systems used by local authorities to set the rents for individual properties. This exercise was undertaken because the Glasgow study found that a number of authorities were in the process of reviewing their rent setting policies in 2001. As Table 2.1 shows, six councils changed their rent setting systems between 2001 and 2006.

Table 2.1: Rent setting systems for Scottish Local Authorities

The definitions of types of rent setting system noted in Table 2.1 follows those used in the Glasgow study. ' GAV'systems are those where rents are in proportion to the 'gross annual values' set for the properties by Regional Assesors in Scotland under the rating system that was abolished in Scotland in 1990. Those values are now very considerably out of date and, as can be seen from the table, a number of the recent changes in policy by local councils have involved the replacement of old GAV systems. However, as of 2006, there were still six councils where rents remained fixed in proportion to GAVs.

'Points' based rent setting systems are now used by 12 councils in Scotland. All but of one the changes introduced over the last five years involved the adoption of a points scheme. Under these schemes, points are allocated on the basis of various characteristics of the dwelling - for example size, type and amenities - and a points total is derived for each dwelling. This provides the basis on which rents are distributed between dwellings in the stock of the landlord.

Points schemes are the predominant form of what are more widely described as 'comparability schemes'. There are, however, other forms of comparability schemes that do not involve the use of points. They are constructed in various ways. Rather than awarding points, 'formula ' schemes set out a range of 'base' rent levels for different types or sizes of dwelling, and those rents are then adjusted to reflect the other characteristics of the dwelling. Those adjustments might take either the form of cash adjustments or percentage adjustments to the base rent. Two authorities in Scotland, East Dunbartonshire and Stirling have schemes of this kind.

Other forms of 'comparability' schemes are neither points nor formula schemes. For example, they might combine elements of those approaches in a hybrid, or simply set out a schedule of rents for a range of property types and sizes.

'Ad hoc' systems are mixed systems that might incorporate any of the above for different parts of their stock. These systems are most typically a legacy of earlier local government re-organisation, where systems that were being operated by individual councils have been merged into a larger structure.

There is a limited relationship between the type of rent setting systems used by local authorities and the degrees of differentiation in rent levels between different types and sizes of dwellings. While all of the councils using GAV systems have relatively wide size-related differentials in their stock (see Appendix C, Table 2), they are not the only ones with wide differentials. Indeed, GAV systems apart, none of the other forms of rent-setting system impose any inherent constraints on the degree of differentiation in the rent levels between one type or size of dwelling and another.

Finally, it may be noted that very few of the rent setting systems used by Scottish councils include location as a factor, whether in points, formula or other systems. There is, however, a locational dimension to GAV-based rents. Consequently, the move away from GAV systems has typically resulted in a reduced incidence in locational variations in rent levels. These issues are considered further in the case studies.

2.5 Service Charges

Information on service charges is far less readily available than information on rents, in respect of both local authorities and housing associations. The information for 15 local authorities which was readily available is shown in Appendix C, Table 5. This indicates that there are some marked differences in accounting practices and that, in some cases, service charges are effectively included within the rent figures shown in the table. In others, disaggregated service charge figures cannot be provided for different types of dwelling.

In those cases where service charge figures are available, primarily for sheltered housing, there are marked variations. Charges range from an average of just £2.64 per week in Argyll and Bute up to an average of £33.48 per week in West Lothian. Service charge figures for flats and houses in the 'general needs' stock were submitted to CIPFA by just 2 councils - Edinburgh and Renfrewshire.

In Edinburgh, the average service charge figure is £1.80 per week, for those flats and maisonettes where a service charge is levied. However, it only applies to 6% of all the flats and maisonettes in the Edinburgh stock. The average service charge for flats and maisonettes in Renfrewshire is higher, at £4.37 per week, and they relate to just less than a fifth of all the flats and maisonettes in the Renfrewshire stock.

There are also limitations on the availability of data on housing association service charges. Disaggregated figures directly comparable to the rent data in Figure 2.2 are not available. Service charge data has been derived from the SCORE (Scottish Continuous Recording System, which monitors RSL settings) returns for lettings made by associations during the course of 2005/06, and these are shown in Appendix C, Table 6.

However, as these figures are based on lettings during the course of the year, they will not necessarily reflect the average position for housing association stock as a whole. Not only do rates differ for different parts of associations existing stock, a substantial part of the lettings data relates to new dwellings entering into the stock for the first time.

The service charge figures in Appendix C, Table 6 are averages spread across all dwellings let during the course of the year; they do not relate solely to those lettings where a service charge was levied.

Service charges are highest for one- and two-apartment dwellings, and these predominantly reflect lettings of sheltered dwellings. For one-apartment lettings, the average service charge was £12.07, and for two-apartment dwellings, £6.63 per week. Service charges for larger dwellings are much lower, averaging just £1.27 per week for three-apartment dwellings, £0.85 per week for four-apartment dwellings, and £0.90 per week for five(+)- apartment dwellings.

Overall, the average service charge for all lettings was £3.90 per week, against an average net rent of £48.10 per week for the same lettings. The average net rent for these lettings is lower than the rents for the stock as a whole (£50.27 - see Figure 2.2), and this reflects the greater proportion of lettings of smaller dwellings in the SCORE data.

It may also be noted that the average service charges levied by the three stock transfer landlords are well below the local averages for other housing associations. While this reflects their local authority origins, and the limited extent to which authorities separate items as service charges, it must also be recognised that there are marked differences in accounting practices between individual landlords within both the local authority and housing associations sectors.

2.6 House prices

For comparative purposes Appendix C, Table 7 shows the levels of average house prices by size of dwelling for each local authority area in Scotland in 2005. Appendix C, Table 8 shows the differentials in house prices by size within each area. These figures exclude values for discounted sales to sitting tenants. They do, however, include values for both new and second hand dwellings. It should also be noted that while this data from the Survey of Mortgage Lenders is uniquely useful in providing house prices by location and size of dwelling, there are issues around the limited sample sizes for some authorities. These issues are particularly acute for the Island authorities and for one-bedroom dwellings.

Figure 2.4 shows the lower quartile house prices for two-bedroom dwellings in 2005 in each local authority area.

Figure 2.4: Lower quartile house prices in 2005 (for two-bedroom dwellings)

For all sizes of dwellings and on three different measures (lowest decile, lower quartile and mean average), house prices are consistently highest in Edinburgh. The sole exception is lowest decile house prices for 4(+)-bedroom dwellings, which are highest in East Lothian.

While a number of authorities consistently have relatively low house prices, the lowest prices vary from one to another according to property size and measure. Angus has the lowest average prices for one-bedroom dwellings; Eilean Siar has the lowest average prices for two-bedroom dwellings; the Shetland Islands has the lowest average prices for three-bedroom dwellings; and Eilean Siar has the lowest prices for four (+)-bedroom dwellings.

For ease of identification, the three authorities with the highest average house prices for each size of dwelling are indicated by dark shading in Appendix C, Table 7, while the three authorities with the lowest average house prices for size of dwelling are indicated by lighter shading.

What is very clear is that the degree of variation in house prices between local authority areas in Scotland is very pronounced. Average prices in the highest area are around three times the level of those in the lowest area, for all sizes of dwellings. The differences for lowest decile and lower quartile prices are similarly pronounced, as can be seen from the lower quartile house prices shown in Figure 2.4 above.

Similarly, Appendix C, Table 8 shows that the range of size differentials for house prices is also generally far greater than for council and housing association rents. One factor is that larger dwellings in the private sector are frequently clustered in particular localities, and the differentials in prices by size may therefore capture an element of price differences based on location.

However, even three-bedroom dwellings have, on average, prices some 44% higher than those for two-bedroom dwellings. In marked contrast, council rents for three-bedroom dwellings are, on average, just 9% higher than those for two-bedroom dwellings, while housing association rents for four-apartment dwellings are on average just 13% higher than those for three-apartment dwellings.

The implications of the very different distributions of house prices (both by local authority area and by size of property) when compared to either council or housing association rents, are considered further below.

2.7 Earnings

Earnings figures, for full time employees for each local authority area are shown in Appendix C, Tables 9 and 10. Table 9 provides earnings data based on the place of work, while Table 10 provides earnings data based on place of residence. In both cases, the data shows lowest decile, second decile, lower quartile, median and mean earnings levels. Data for the lowest decile measure is not available for the Island authorities due to small samples and, consequently, high standard errors.

Again, it is evident that there are substantial variations in earnings across Scotland, albeit far less marked than the variations in house prices. There is also rather greater variation in earnings based on place of residence, rather than based on place of work. However, those variations are greater for average (and above) earnings than for below average earnings, reflecting the tendency for longer journeys to work by those with higher earnings. The limited extent of local variations in lower quartile place of work based earnings can be clearly seen in Figure 2.5.

Figure 2.5: Lower quartile earnings in 2005 (place of work based earnings)

On the basis of place of work, earnings in the highest earning area in Scotland (Aberdeen City) are less than 25% higher than in the lowest earning area (Western Isles), for the lowest and second decile cases. Even in the lower quartile, the differential between the (same) highest and lowest areas only rises to 34%. In contrast, the differential for average earnings is just under 50% on the basis of place of work, and just over 50% on the basis of place of residence.

In broad terms, the degree of geographical dispersion of earnings levels for low earners is of the same order as the dispersion of housing association rents across Scotland (if the LSVT landlords are excluded), but far less marked than the dispersal of local authority rents.

2.8 Contrasting differentials

Appendix C, Table 11 provides a summary of the degree of geographical dispersion of rents, house prices and earnings. As well as showing the different ranges of dispersion, it also shows the rankings for each local authority area for each measure.

From a simple glance at the table, it is apparent that there is very little correspondence in the rankings for rents and the rankings for either house prices or earnings. The exception is Edinburgh, which is either top or very close to the top on all measures.

While some care is needed in terms of the data for the island authorities (due to small sample sizes for the house price and earnings data), nonetheless there are some remarkable divergences in the rankings on different measures for some authorities.

Dundee City has relative high council rents, but very low house prices and only moderate earnings. Inverclyde has relatively high local authority and housing association rents, but only moderate house prices and low earnings. Highland has a relatively high ranking for council rents and a low ranking for housing association rents, but moderate house prices and earnings.

East Lothian has low council rents, but very high earnings, house prices and housing association rents. East Renfrewshire has high rents and house prices, but very low earnings. West Lothian has very low housing association rents, but moderate house prices and relatively high earnings and council rents.

If it is clear that the greatest anomalies arise in connection with council rents, it is also notable that the profiles of house prices and low place of work earnings are also very divergent. This conflict between key housing and labour market indicators has important implications for the construction of a national rent policy.

2.9 Affordability

An initial analysis of the affordability of local rents in the social rented sector in 2005/06 is provided in Appendix C, Table 12. A more detailed analysis of the affordability of social sector rents is to be found in Chapter 4, which considers a range of options for the future of rent policy in Scotland compared to the 'baseline' of existing 2005/06 rents.

The initial analysis in the next chapter is based on the rents for three-bedroom (or four-apartment) dwellings. Appendix C, Table 12 shows gross rent to gross earnings ratios for a full time earner at the level of the minimum wage, based on a 40 hour week. Against this earnings level, gross rent-to-earnings ratios range from 19.6% to 31.0% in the local authority sector, and from 27.4% to 31.9% in the traditional housing association sector. The highest ratio, however, is 33.0% in the case of Glasgow Housing Association ( GHA).

The table also shows that with rents and earnings at those levels, a couple with 2 children would be dependent on housing benefit only in a small number of areas with relatively high rents: Edinburgh in respect of council rents; Glasgow in respect of GHA rents; and Aberdeen City, East Lothian, Edinburgh, Inverclyde, Shetland Islands and South Ayrshire in respect of housing association rents.

Even in those areas however, the extent of housing benefit dependency is very limited. With an earned income of £194 per week (based on 40 hours at the minimum hourly wage of £4.85 per week in 2005) a couple with 2 children would only qualify for housing benefit where rents exceed £60.14 per week. For a three-bedroom dwelling, the highest average weekly rent in the sector in 2005/06 was £63.96 in the case of GHA; less than £4 over the level where the minimum wage household qualifies for housing benefit.

There are also affordability issues for pensioner households, and these are considered in more detail in Appendix B. There is incomplete data on the distributions of incomes of pensioner households in the social rented sector, but an analysis of that limited data does suggest that some four fifths of all pensioner households in the sector are already in receipt of housing benefit. With the current structure of the UK-wide housing benefit scheme, any changes in rent levels would result primarily in changes to the level of housing benefit those households receive; only in a minority of cases would such changes have a direct impact on the disposable income of pensioner households. The issues for pensioner households are also considered further in the following chapter.

2.10 Comparing mortgage costs and rents

As seen above, house prices vary very substantially from area to area and have a very different geographical profile than that for council and housing association rents. This has a number of implications for rent policies, and these are considered in more detail in the following chapter.

However, Appendix C, Table 13 makes a simple comparison between, on the one hand, the costs of house purchase at the low end of the local housing market (lowest decile prices), and on the other, average local housing association and council rents for three-bedroom (or four-apartment) dwellings. This shows the rents as a percentage of the simple costs of a 100% mortgage at the average interest rate prevailing in 2005 (5.2%). Thus, the measure neither makes any provision for either deposits or for repair and insurance costs, although inevitably those items will offset each other to some degree.

However, it is notable that on this basis there are two areas where mortgage costs would be lower than local council or housing association rents. These cases, shown in bold in Appendix C, Table 13, are Eilean Siar and the Shetland Islands. It should be stressed, however, that the results for the island authorities need to be treated with some caution due to small sample sizes for house price data.

In addition, there are a number of areas where rents are very close to mortgage costs. In Dundee City, both council and housing association rents are at levels equivalent to more than 85% of mortgage costs, as are housing association rents in East Ayrshire.

The proximity and overlap between social sector rents and the costs of house purchase at the lower end of the local housing market raise a number of issues. The first is one of equity: that is to say, is it equitable that tenants should be required to pay a rent that is close to or higher than the costs of local house purchase for the use and occupation of a social sector dwelling?

The second issue is about tenure polarization. If it is little or no more expensive to buy than to rent, there are clear incentives for tenants to consider purchase. In the council sector this could mean tenants are effectively incentivised to exercise the right to buy because, particularly when discounts are taken into account, this may be cheaper than renting (albeit before taking account of repairing responsibilities). In other cases, it will provide an incentive for working households to move out of the social sector in order to buy, further contributing to the processes of social and neighbourhood polarization.

2.11 Concluding remarks

This chapter has provided an analysis of the profile of council and housing association rents in 2005/06, and their relationship to earnings and house prices across Scotland. The findings in this chapter are elaborated further in Chapter 3, which examines rent levels within the case study landlords. This is of particular significance in providing more detailed information on the distribution of rents levied by individual landlords.