Cities continue to grapple with a rising demand for housing, which affects affordability and the well-being of its citizens. This growth continues to put pressure on the delivery of adequate, affordable housing in well-located areas while the availability of infrastructure and proximity to economic nodes remains a challenge. This has led to increasing infill development of medium-density to high-density affordable housing in greenfield areas located adjacent to higher-income neighbourhoods.
This study investigates how a new affordable housing development influences the locational and structural values of the adjacent, existing housing market.
Transactional data of residential sales for two areas in South Africa are used to measure the value change. Both areas are located within an urban setting next to an open, greenfield area that was redeveloped for affordable housing.
Two case studies are used and analysed with hedonic pricing modelling to identify and measure the value change for the locational and structural characteristics before and after the development of affordable housing.
The results reveal a changing housing market as the locational and structural characteristics change in value, further highlighting the importance of careful planning that preserves the existing market and also supplies affordable housing.
The value of several structural characteristics of properties will change, revealing just how consumer preference responds when affordable housing is introduced in an existing housing market. Distance to an affordable housing project continues to influence the house market value and careful consideration should be made when planning to integrate an affordable housing development in an existing neighbourhood.
The urban settlement patterns in South Africa have transformed from the post-apartheid era; however, cities continue to grapple with increasing demand for housing within its boundaries. This has led to the promoting of residential density by encouraging integrated mixed-use development (ed. Smith 2003:2), as well as utilising areas that have existing or easy access to horizontal infrastructure. Since 1994, various approaches were undertaken by the government to support the increasing demand for housing and since property was available and more reasonably priced at the edge of urban areas, most affordable housing developments took place at the periphery (Schoonraad
Housing affordability within the global economy is emerging as a crisis, as housing costs and household income continue to mismatch (Wetzstein
In a recent article, Phakgadi (
This article proceeds as follows. First, we discuss the regulating environment related to affordable housing development in South Africa, as well as the extant literature that relates to affordable housing provision and its effect on the surrounding built environment. This is followed by the methodology and results while we end with a conclusion.
Recent news media continue to highlight the nature of urbanisation where housing affordability is at the forefront of the debate (Leshage
In 2004, the National Department of Housing (NDoH) introduced the Breaking New Ground (BNG) policy in an effort to provide sustainable human settlements and increase the delivery of appropriate housing in the urban core of South African cities. This was done to address the frequent mismatch of housing supply and demand at well-located and in-demand areas. The Integrated Residential Development Programme (IRDP) within the BNG policy was introduced as a result.
The purpose of the IRDP was to foster integrated housing development that caters for a broad spectrum of typologies and affordability levels within the housing market. The programme delivers four different housing typologies, including government subsidy housing, social housing, a finance-linked individual subsidy programme (FLISP) and bonded segment housing. Affordable housing can be termed as a house provided by a social housing government institution or an accredited social housing project, built in a designated restructuring zone for low-to-middle-income earners (Social Housing Policy for South Africa
The FLISP housing is intended to assist households to access housing, by providing partially subsidised housing that requires a lower deposit on a house. Individuals who can afford personal loans up to R300 000 are eligible for FLISP housing. FLISP housing is greater than 40 m2 and characterised as a detached, semi-detached single or double-storey building (see
Affordable housing. (a) Semi-detached single-story house. (b) Medium-density social housing.
Social housing, illustrated in
Many of the programmes that were implemented by the government to combat the housing backlog, provided housing at a sub-optimal location, often located on the outskirts of urban cities and disconnected from economic networks, hindering sustainable livelihoods and opportunities for residents (COGTA
The property price or market price is the actual price of the property on which the consumer and the buyer agree (Pirounakis 2013:384). These values are valued differently by buyers and sellers and also based on the characteristics of the house (Ball, Lizieri & MacGregor
The location forms part of the value proposition by the market. Tiebout’s (
The location factor that is considered when choosing a residential property reflects the individual’s preferences and choice of the surrounding neighbourhood. It has an impact on the household’s well-being and quality of life (Uchenna
Several studies (Du Preez & Sale
With the migration to large urban areas of people seeking employment, the increasing urban population drives demand for housing which regularly outpaces the supply, especially in the affordable housing segment (Blaauw et al.
Pirounakis (2013:27) describes property price as the value generated from the actual price of the property both the consumer and seller agree upon when making a property transaction deal. This definition assumes both the seller and the buyer have sufficient knowledge of the property and the property market. Residential property is valued as a heterogeneous product, which comprises a bundle of inherent attributes or characteristics that may not be separated from each other since these components refer to the implicit price of the property (Woo
The characteristics of a house can be divided into structural characteristics, that is, the physical appearance of a property, including the number of bedrooms, number of bathrooms, property age and erf size and its immediate surroundings, as well as locational characteristics, that is, the location unique to a house, proximity to police stations, schools, clinics and retail centres (Goodman
Sirmans, Macpherson and Zietz (
The top 20 characteristics.
Variable | Appearances | Times positive | Times negative | Times not significant |
---|---|---|---|---|
Age | 78 | 7 | 63 | 8 |
Time on the market | 18 | 1 | 8 | 9 |
Lot size | 52 | 45 | 0 | 7 |
Ln lot size | 12 | 9 | 0 | 3 |
Square feet | 69 | 62 | 4 | 3 |
Ln square feet | 12 | 12 | 0 | 0 |
Brick | 13 | 9 | 0 | 4 |
Fireplace | 57 | 43 | 3 | 11 |
Basement | 21 | 15 | 1 | 5 |
Air-conditioning | 37 | 34 | 1 | 2 |
Garage spaces | 61 | 48 | 0 | 13 |
Deck | 12 | 10 | 0 | 2 |
Pool | 31 | 27 | 0 | 4 |
Bedrooms | 40 | 21 | 9 | 10 |
Number of stories | 13 | 4 | 7 | 2 |
Number of bathrooms | 40 | 34 | 1 | 5 |
Full baths | 37 | 31 | 1 | 5 |
Number of rooms | 14 | 10 | 1 | 3 |
Distance | 15 | 5 | 5 | 5 |
Time trend | 13 | 2 | 3 | 8 |
Property owners have a common belief that affordable housing development located near their homes will automatically decrease their property’s value and the neighbourhood’s aesthetic qualities. The latter is termed the not-in-my-backyard (NIMBY) theory and based on the idea that affordable housing will be visually unattractive, poorly maintained and managed, which will also, in turn, increase traffic and the level of crime in an area (Habitant & Humanity
A study compiled by Du Preez and Sale (
The position of this article is to build on the existing literature by identifying if such a development results in a change in market value of the structural and location characteristics of the existing neighbourhood with the development of an affordable housing project. This would enable improved decision-making when identifying open spaces for such developments to encourage appropriate development that aligns with the existing market of the adjacent neighbourhood.
This study follows a combined research approach through a case study analysis and a hedonic pricing to determine the change. A cross-sectional data series of various locational and structural aspects of residential properties is used to investigate the residential market structure pre- and post-development. Pre-development is defined as the period leading up to, but excluding, the construction phase of the affordable housing development. Post-development refers to the period once construction starts, and was chosen since price shifts were observed in the market once construction started, coupled with long construction times on large residential projects. The market responded immediately in both case studies, a suggestion that the NIMBY theory applies earlier than would generally be expected. This is encapsulated by the public’s perception of affordable housing developments in close proximity to existing neighbourhoods as detrimental to the area, more specifically, raising the concern about an upsurge in crime due to construction, overcrowding and the general disturbance of peace (Du Preez & Sale
The two case studies selected are in South Africa: Fleurhof in Randburg and Birch Acres in Kempton Park. These developments were chosen since both case studies are embedded in a previously open-space context where affordable housing typologies have been developed next to an existing neighbourhood, characterised by bonded (middle-to-high-income) housing. Since the residential developments in both case studies have already been completed, they are ideal to analyse and develop policy implications.
The hedonic pricing model is used to value the property’s unique attributes and combine these to estimate the price of a property. In doing so, it is assumed that individual households have full information concerning the price and the unique attributes of the residential property (Ham, Maddison & Elliot
The hedonic pricing model equation states that the market price of a property is expressed as a function of the structural and locational characteristics (Woo
In these equations:
The residential property data on property prices, structural characteristics and locational characteristics are quantifiable variables. The data includes numerical information such as the actual property price, house attributes and measurable distances. The model used to determine the real price of a property is given as:
The dependent variable in the hedonic pricing model, Real_Price, represents the individual property sales price, β1 – β7 represents the independent variables or structural attributes of each property and Z1 represents the distance between the individual property and the open space (pre-development), as well as the proximity between the property and the affordable housing development (post-development). The property prices were adjusted for inflation.
The residential property data was sourced from Property 24® and Lightstone Property. Du Preez and Sale (
This article followed all ethical standards for research without direct contact with human or animal subjects.
Fleurhof originally consisted of mainly low-density, single dwelling freehold residential units. The affordable housing development utilises the greenfield area adjacent to the neighbourhood and developed high-density freestanding and semi-detached housing, providing for an estimated 83 000 people (Calgro M3
The integrated residential development at Fleurhof started construction in 2013 and continued to be under construction in 2018. The period between 2000 and 2012 is considered the pre-development period with 124 transactions while the post-development period is between 2013 and 2018 with 68 transactions (see
Fleurhof, pre- and post-development.
A comparison between the average house price before the development with the proposed development price range reveals a significant difference. In 2012, the average property price for Fleurhof was R583 845 (adjusted for inflation) while the suburb had an average of three bedrooms and two bathrooms per house, with an average erf size of 848 m2. This value of the affordable housing project is more than 50% below the average house price for the neighbourhood prior to its development. This reveals a significant change in the type of residential typology that is introduced to the market. Keeping this price differential in mind, the results of the hedonic pricing model for both periods are summarised in
Fleurhof model summary.
Model | Unstandardised coefficients |
Significance | ||
---|---|---|---|---|
Standard error | ||||
(Constant) | −38 445 424 | 11 871 627 | −3.24 | 0.00 |
Year | 19 361 | 5913 | 3.27 | 0.00 |
Bed | 51 395 | 24 166 | 2.13 | 0.04 |
Bath | 19 456 | 28 982 | 0.67 | 0.50 |
Garage | 63 963 | 23 566 | 2.71 | 0.01 |
Erf | −86 | 106 | −0.81 | 0.42 |
Pool | −13 | 2826 | −0.004 | 0.996 |
PROP_AGE | 116 804 | 57 220 | 2.041 | 0.043 |
Distance | −23 | 177 | −0.13 | 0.89 |
(Constant) | 35 519 420 | 32 403 110 | 1.10 | 0.28 |
Year | −17 025 | 16 074 | −1.06 | 0.29 |
Bed | −53 195 | 32 718 | −1.63 | 0.11 |
Bath | 78 544 | 33 601 | 2.34 | 0.02 |
Garage | 13 625 | 37 927 | 0.36 | 0.72 |
Erf | −134 | 137 | −0.98 | 0.33 |
Pool | 15 273 | 62 133 | 0.25 | 0.81 |
PROP_AGE | 350 | 4115 | 0.09 | 0.93 |
Distance | −782 | 262 | −2.98 | 0.00 |
Fleurhof model summary.
Model | Adjusted |
Standard error of the estimate | Significance | ||
---|---|---|---|---|---|
Fleurhof pre- development | 0.232 | 0.179 | 216487.019 | 4.357 | 0.000* |
Fleurhof post- development | 0.281 | 0.184 | 214974.912 | 2.885 | 0.009* |
Significant on
Comparing the results of the pre-development and post-development model reveals a change in several variables and while these changes will be discussed in more detail in the following section, a couple of relevant changes are emphasised. The prominent result from the model in
In most instances, a positive coefficient is expected as bedrooms increase with the price of a house as highlighted by Dodds (
The distance variable continued to be negative in the post-development environment, albeit higher. This suggests that the affordable housing typology introduced to the area did not affect the prices of nearby houses as expected, although priced below the average suburb price. It should be noted that even though the distance is statistically significant, the number of property sales close to the project is low when compared to the total transactional value, with the majority observed farther away from the project.
For this reason, an analysis of the transaction data for homes within 160 m of the affordable development revealed that there is indeed a positive relationship between distance and property prices in Fleurhof. This result aligns with the existing literature in South Africa that finds similar results. The results hold for both the pre-development and post-development scenarios and are illustrated in
Price and distance relationship. (a) Distance up to 160 m, pre-development. (b) Distance up to 140 m, post-development.
Birch Acres located to the north of Kempton Park is a large suburb with a variety of residential typologies. In 2003, prior to the affordable housing development, the average house price for properties located within a 400-m buffer of the greenfield area was R190 857 (adjusted for inflation). The suburb had an average of three bedrooms and two bathrooms per house, while the average erf size was 924 m2.
The development in Birch Acres was aimed at the affordable housing market, particularly the FLISP market and units sold for an average of R74 000, at 61% lower than the buffer area average house price (Demacon
Birch Acres, pre- and post-development.
Birch Acres model summary.
Model | Unstandardised coefficients |
Significance | ||
---|---|---|---|---|
Standard error | ||||
(Constant) | −7 029 546 | 7 155 639 | −0.982 | 0.327 |
Year | 11 560 | 6788 | 1.703 | 0.091 |
Bed | 6045 | 6154 | 0.982 | 0.327 |
Bath | 14 253 | 4202 | 3.392 | 0.001 |
Garage | 103 | 39 | 2.596 | 0.01 |
Erf | 39 303 | 15 548 | 2.528 | 0.012 |
PROP_AGE | 2543 | 896 | 2.836 | 0.005 |
Pool | 34 813 | 3572 | 0.974 | 0.331 |
Distance | 98 | 43 | 2.259 | 0.025 |
(Constant) | −338 613 788 | 13 992 142 | −2.42 | 0.016 |
Year | 16 889 | 6974 | 2.421 | 0.016 |
Bed | 4620 | 21 394 | 0.216 | 0.829 |
Bath | 13 953 | 17 510 | 0.797 | 0.427 |
Garage | 11 797 | 11 155 | 1.058 | 0.292 |
Erf | 209 | 77 | 2.709 | 0.007 |
PROP_AGE | 5460 | 1988 | 2.746 | 0.007 |
Pool | 62 428 | 40 804 | 1.53 | 0.128 |
Distance | 354 | 117 | 3.02 | 0.003 |
Birch Acres model summary.
Model | Adjusted |
Standard error of the estimate | Significance | ||
---|---|---|---|---|---|
0.286 | 0.25 | 51801.187 | 7.962 | 0.000* | |
0.339 | 0.31 | 144859.33 | 11.773 | 0.000* |
A comparison of the pre-development results with the post-development results reveals that the bed and bathroom variables have stayed relatively consistent in both periods. The results indicate that properties located near an open space (pre-development) were slightly negatively affected by the proximity to the open field as property prices would increase by R10 per metre or R1000 per 100 m as distance increased away from the greenfield area. Post development, the value is significantly higher and increased by R35 400 for every 100 m further away. This result aligns with previous studies of a similar nature in South Africa (Du Preez & Sale
Both case studies confirm that there are changes to how to the locational and structural characteristics of the adjacent existing property market are valued before and during the construction of an affordable housing project. The results reveal that in both case studies, the property age, number of bedrooms, bathrooms and garage spaces is positively affecting the price of the property before the development.
This continued to be positive in Birch Acres, post-development; however, in Fleurhof, a change is observed in bedrooms with a negative coefficient. As explained earlier, this change could be a result of changing consumer preference within the residential market, possibly due to the size and target market related to the housing affordability development.
The erf size changes between the two periods are similar in both case studies. In Fleurhof the negative coefficient increased in the post-development model, suggesting that smaller is preferred by the market and more so after the development. In Birch Acres the value decreased substantially, and although it remained a positive coefficient, the change also suggests that the market moved to smaller erf sizes, aligning with the affordable housing project provisions.
The coefficient for garages remained positive in both case studies and in both models, with Fleurhof showing a lower value associated with an increase in garages, while in Birch Acres more value is associated with increasing garages. A comparison of the census data in 2001 and 2011 (Quantec
The availability of a pool has increased in value in both case studies after the development. This result is unexpected and could likely reveal an important structural characteristic that supports a property’s value, irrespective of changing market conditions.
The locational characteristic of distance reveals a negative coefficient in both periods for Fleurhof but increasing post- development. This suggests that properties closer to the affordable development are valued more than those further away. However, an additional analysis presented earlier reveals that the opposite is revealed for properties near the greenfield site pre- and post-development, with prices increasing as distance from the development site increased up to 140 m post development. In Birch Acres the coefficient continues to be positive and increase, aligning with the existing literature that locating closer to affordable housing has a negative effect on price. In both case studies the value of the distance coefficient increases in the post-development period.
The hedonic pricing model during both periods reveals the change in the structural and locational characteristics that influence the market price of houses in each case study. Additional to these observed value changes is the price alignment between the existing market and the newly proposed development. In other words, the price differential.
This suggests an analysis of the original price difference between the average property price for the existing neighbourhood versus the average sales price of the affordable housing project. The average sales price in Fleurfof prior to the development was R584 000 (inflation-adjusted), while the affordable housing average price was R280 000 (inflation-adjusted) which translates into a price difference of 52%. At Birch Acres, the average house price was R191 000 (inflation-adjusted) while the average price at the affordable housing development was R70 000 (inflation-adjusted), a 61% difference. The higher price differential in Birch Areas could further support a positive and increasing distance coefficient evident at Birch Acres, but not at Fleurhof. This points towards a potential price differential cut-off point where a new development between 52% and 61% from the average price has a negative effect on the existing house prices of the neighbourhood.
With the increasing demand for land to develop affordable housing development, coupled with the pressure to use infill, greenfield development to limit urban sprawl and utilise existing infrastructure, the results from this study provide important insights to local and regional councils that need to deliberate on relevant planning aspects.
The current market pricing, locational and structural characteristics need to form part of a market analysis that evaluates how an affordable housing development could support, rather than negatively influence the existing adjacent housing market. Ideally, the affordable housing development should reflect comparable characteristics to the surrounding area.
Delivering adequate and affordable housing for a growing market in well-located areas has resulted in the development of affordable housing projects in open spaces next to established middle-to-high-income neighbourhoods. This is in response to limited available land on the periphery of the urban edge, as well as supplying affordable housing in areas located closer to economic nodes that provide reasonable employment opportunities. There is a growing concern in the marketplace about the effect that such developments have on the existing adjacent residential property prices. In an effort to improve decision-making of the regulatory environment, as well as expand on the knowledge of affordable housing and its effect on the existing market, this study provided insight into that and should be useful for both local and regional planning departments when faced with similar applications. The analysis of case studies provides useful support in examining what has worked and what could be done differently.
The results from the hedonic pricing modelling indicated that structural and locational characteristics of the neighbourhood changed in value after the development of affordable housing. The change in value for the structural characteristics was an unexpected result and shows that consumer preferences could alter the ideal typology of the neighbourhood housing market. Distance to an affordable housing project continues to influence the house market value and careful consideration should be made when planning to integrate an affordable housing development into an existing neighbourhood. Ideally, the market price of these units should align with the current market and, to this effect, a price differential analysis could provide useful guidance. Expanding on the market differential analysis is a topic that would require further research.
Lastly, the affordable housing development could consider preserving a part of the greenfield area to represent a green buffer with playgrounds between the existing neighbourhood and affordable housing project. Furthermore, the layout of the housing typology should consider and position the higher-priced affordable units closer to the existing neighbourhood to support the existing market values.
The authors thank FEMS for their financial contribution; NWU for the great opportunity to study and Christien Terblanche for the language editing.
The authors have declared that no competing interest exists.
All authors contributed equally to this work.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Data sharing is not applicable to this article as no new data were created or analysed in this study.
The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of any affiliated agency of the authors.