Urban Studies, Vol. 40, Nos 5–6, 897–916, 2003

Markets and the Structure of the Housebuilding Industry: An International Perspective

Michael Ball

[Paper first received, November 2002; in final form, December 2002]

Summary. This paper sets out to examine why the modern housebuilding industry is organised in the ways it is and to identify some implications for the wider operation of housing markets. It concentrates on advanced economies and the impacts of market conditions, regulatory con- straints, production characteristics, institutional structures and land supply. Widespread differ- ences occur across countries in the ways in which housebuilding is institutionally structured. It is argued that these differences are generally explicable in economic terms and that regulatory practices determine much of the variation.

Introduction This paper sets out to examine why the mod- neatly into the standard economics lexicon as ern housebuilding industry is organised in a competitive industry. The well-known the ways it is and to identify some implica- results for a competitive industry from tions for the wider operation of housing mar- the structure–conduct–performance paradigm kets. It concentrates on advanced economies would, therefore, seem to be reasonable hy- and the impacts of market conditions, regula- potheses to apply to housebuilding: namely, tory constraints, production characteristics, that long-run risk-adjusted returns are low institutional structures and land supply. and that housebuilding costs closely reflect At first sight, the discussion seems simple the marginal costs of inputs. In a dynamic and straightforward. Housebuilding is char- framework, there is a cyclical effect with acterised by the existence of a large number supply constraints pushing up builders’ mar- of relatively small firms. They also use pro- gins and their input costs in housing market duction techniques that are labour-intensive booms and depressing them in slumps. Yet, and change relatively slowly. They also are these temporary variations should be limited, ‘flexible’ in order to be able to adapt to given the ease of entry and exit. These con- potentially large variations in output. Gener- clusions make life easy for housing market ally, as a result, scale economies are low, modelling, as the supply side can be dealt which helps both to make entry and exit from with simply by reference to input cost indices the industry relatively easy and to explain the and a cyclical mark-up to reflect changes in small average firm size. builders’ profits. Housing supply, consequently, seems to fit Several characteristics of real-world

Michael Ball is in the Department of Real Estate and Planning, University of Reading, Whiteknights, PO Box 217, Reading RG6 6AH, UK. E-mail: [email protected].

0042-0980 Print/1360-063X On-line/03/05/60897–20  2003 The Editors of Urban Studies DOI: 10.1080/0042098032000074236 Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 898 MICHAEL BALL housebuilding, nevertheless, seem to be at they cannot be the full answer to the compet- variance with the simple competitive model. itive puzzle. Housebuilding market contexts, First, housebuilding takes on a number of prevailing technologies, input costs and input different institutional forms across and availability need to be explored in greater within countries. There is remarkably little depth. international information, or data, on the na- ture of housebuilding industries in the oceans Housing Output Variations of housing-related international information, but sufficient exists to know that housebuild- Housebuilding is well known as a cyclical ing industrial structures vary quite consider- industry with variations in output and prices ably. Secondly, firms adopt strategies and, driven by cycles in the housing market as a from experience, know that they are import- whole. Such volatility is typical of a con- ant in determining profit outcomes. Further sumer durable or investment goods industry features could be elaborated, but these two where the existing stock of a good is far are sufficient to highlight the puzzle. greater than current output of it. New hous- Namely, how are these characteristics poss- ing represents only a relatively small per- ible, if the simple competitive market model centage of all the houses bought and sold in is correct? With competition, there should be mature housing markets. In the UK, for ex- strong pressures to create similar types of ample, new building averages only around 15 firms rather than the observed international per cent a year of total owner-occupied house diversity. Furthermore, strategic behaviour sales. With such low ratios of new output to can have no effect on market outcomes in the the existing stock, changes in overall housing purely competitive model (Martin, 1994). demand tend to induce large variations in the The competitive model, none the less, demand for newly built housing. This is be- would predict some industry differences. As cause new building is the prime source of housebuilding techniques are widely known, extra housing to accommodate any increases it would suggest that housebuilding indus- in overall housing demand. It takes a long tries vary primarily due to differences in time, moreover, to build sufficient houses to relative input prices between countries. For satisfy demand increases, so that when de- example, high-wage countries should adopt mand grows, prices rise in response to short- less labour intensive techniques and, so, have ages and overshoot their long-run production processes reflecting that fact— equilibrium levels as households try to find with, for example, a greater use of off-site homes before supply has caught up. fabrication and on-site equipment. Such Housing market cycles are an inevitable choice-of-technique influences help to deter- consequence, therefore, of housing supply mine the character of housebuilding indus- mechanisms (Ball and Morrison, 2000). The tries between countries. For example, extent to which volatility occurs more in high-wage, timber-rich North America and prices or in levels of housebuilding depends Scandinavia have housebuilding industries on the extent to which extra housing can be relying on large-scale suppliers of pre-manu- supplied when prices rise and the speed of factured timber-frame systems and exten- that response. sively use pre-assembled modules of internal One reason why house prices fluctuate house fittings. In contrast, high-wage but more than the prices of many other durable timber-poor, the Netherlands and France goods arises from the structure of housing more commonly adopted concrete panel off- markets. In industries like cars and domestic site fabrication for house structures. Such appliances, which also face similar ‘stock- differences affect labour skills mixes, adjustment’ demand characteristics, econom- amongst other things, but they do not negate ies of scale in production tend to generate the general hypothesis of small, flexible oligopolistic industrial structures. These en- building firms with no strategic options, so able firms to hold prices relatively steady

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 899 during downturns in demand. They prefer to cussed above. However, there are several key accept the resultant possibly greater demand features which have longer impacts. fluctuations in volumes in order to sustain the They include variations in household num- margins necessary to finance their large bers generated by population and social amounts of fixed capital. Such strategies are changes; migration into prosperous regions; not realistic in housebuilding, because of the technologies of living (the tram/streetcar, the ease of entry and exit and the competition in bus and the motor car being the most histori- the housing market from existing housing cally notable ones); previous shortages owners. This means that, in downturns, there caused by catastrophic events like war; and are always large numbers of suppliers pre- public policy towards such matters as subsi- pared to cut prices for a quick sale when dies and tax-breaks for particular types of financial pressures dictate. housing. These when combined together tend This pricing characteristic of house- to create housebuilding ‘long’ cycles that are building is ignored by those who, over the superimposed on the business cycle. Many of decades, have argued that housing pro- these longer-term factors are extended lag duction should become far more capital- responses to specific demographic, techno- intensive, like a ‘real’ consumer goods logical or policy ‘shock’ events, which industry. Unless supported by huge public makes it difficult to generalise about house- subsidies, housebuilding firms that attempt to building long cycles, though many have become highly capital-intensive in their pro- tried. They tend, however, to be important duction methods almost certainly go bust influences on the structure of contemporary during the next major downturn. Despite this housebuilding industries through their effects characteristic of housebuilding, huge experi- on the overall level and type of housing ments in manipulating housebuilding de- demand. mand to smooth output fluctuations were The short and long cyclical behaviour just undertaken in a number of European coun- described can be seen in the graphs in Figure tries from the 1950s to the 1970s by taking 1ofthe share of housing investment in na- advantage of social housebuilding pro- tional income from 1956 to 2000 for seven grammes (Donnison, 1967; Barlow and advanced countries: Canada, France, Ger- Duncan, 1994; Harloe, 1995). Large house- many, Japan, the Netherlands, the UK and builders emerged and governments tried to the US. The impact of the business cycle can create the market conditions that would en- be seen in all the country data, although it is able them to reap cost savings through scale more marked in some countries than others. economies and innovation. Unfortunately, It shows up because of the way housebuild- few of the scale economies in production ing varies disproportionately with changes in materialised and most of the large house- housing demand, encouraging the rises as a builders that grew up disappeared as soon as share of national income during business cy- public funding ceased, as did the innovatory cle upswings and falls during downturns. At building systems they sometimes used (Ball, the same time, however, long-cycle behav- 1988; Glendinning and Muthesius, 1994; iour can also be seen. Gann, 1996). Figure 1 also shows that housebuilding Some causes of changes in housing de- varies in importance between countries over mand are short-term, business-cycle-related very long periods of time. This characteristic factors, like rising incomes and interest rates. is summarised in the data provided in Table So, unsurprisingly, housebuilding tends to 1, where countries are ranked by the long- fluctuate in line with the business cycle; with term significance of housebuilding as a share the degree of output and price variation being of GDP. The country with the most house- greater than that for the rest of the economy building over the past half a century has been on average, because of the stock adjustment Germany, followed quite closely by France, characteristics of new housing demand, dis- Japan, the Netherlands and Canada. The US

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 900 MICHAEL BALL

Figure 1. Fixed residential investment as a share of GDP, 1956–2000. Source: OECD, INSEE.

trails almost 2 per cent behind Germany and lower-income groups. It has been argued that the lowest, the UK, is behind by almost 3 per housing investment declines as a share of cent. These differences can be explained by national income as national income rises several factors. Germany is at the highest above a certain per capita level, but the rank because of the housing boom during the long-cycle characteristics of these data make period after reunification in the 1990s. The it hard to isolate any such effect (Malpezzi, US provides no subsidised low-income hous- 1990). ing, unlike the European countries listed, Table 1 also suggests that there are maxi- which probably accounts for most of the mum amounts of housing investment that a difference between it and them. The UK is well-functioning economy can cope with. It exceptional, probably because of its restric- shows that housing investment shares have ted land supply in growth regions and a never risen above 7–8 per cent of GDP for correspondingly high propensity to repair the countries listed, even when there were and upgrade existing dwellings. chronic shortages. This is unsurprising as Once these differences are taken account such percentage figures imply a huge alloca- of, a crude estimate of the typical share of tion of a country’s existing resources into housing investment in the annual national housebuilding. Not only does this deny those income of an advanced economy with mod- resources to other activities, but it also erately expanding household numbers would necessitates substantial knock-on invest- seem to be around 4–5 per cent, with the ments in building materials industries, the extra percentage point depending on the so- training of skilled labour and urban infra- cial attitude to the housing conditions of structure. When the housing boom is over,

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 901

Table 1. Housing investment as a share of national income, 1956–2000: cross-country comparisons

Mean Maximum Minimum Standard deviation Standard deviation/Mean

Germany 6.4 8.1 4.8 0.9 0.15 France 5.7 7.8 4.1 1.2 0.20 Japan 5.6 8.8 3.6 1.3 0.23 Netherlands 5.3 6.4 3.9 0.6 0.12 Canada 5.3 7.4 3.8 1.0 0.19 USA 4.4 5.7 3.2 0.6 0.14 UK 3.5 4.7 2.6 0.5 0.14

Source: OECD. furthermore, all those resources have to be mation on construction, land and housing redirected elsewhere again, which can gener- markets militate against it. ate substantial re-adjustment costs—as Ger- Long-run housing investment data, in ad- many has experienced in recent years. These dition, help to identify the fact that house- factors highlight the difficulties of expanding building has been of declining relative housing supply rapidly to meet sharp in- economic importance since the 1980s in creases in demand. Canada, France, Japan and the UK and, since The degree of volatility in housing invest- the mid 1990s, in Germany. Only the Nether- ment GDP shares can be measured by look- lands and the US saw sustained expansion ing at the standard deviations of annual throughout the 1990s, fuelled by their re- residential investment shares in national in- spective economic booms and household come, as shown in Table 1, as well as growth at the time. through visual inspection of the graphs of House prices, as well as output levels, Figure 1. Typically, housing investment exhibit marked cyclical patterns, as can be fluctuates annually by around 0.5–1 per cent seen in data on annual real house price of GDP, with the countries with the largest changes for France, the UK and the US from shares of housing investment in national in- 1985 to 2000, shown in Figure 2. All three come, not surprisingly, experiencing the countries experienced several years of falling largest absolute fluctuations in those shares. real house prices in the early 1990s, follow- When weighted by their respective means, ing buoyant, or even bubble, price rises in however, the standard deviations across the second half of the 1980s. House prices countries are quite similar. were again on the rise in the late 1990s. Even Housebuilding fluctuations may vary greater variations occurred in regions of markedly over time, but overall housebuild- these countries, and housebuilding tends to ing fluctuations are not atypically large when be a regional business. Housebuilders, conse- compared with other capital goods industries, quently, face uncertain future prices for their which experience significant output products. fluctuations as well (Ball, 1996). What is also What do these general market data imply interesting is that the cycles of housing in- for housebuilders? First, they highlight that vestment in each country in the past have housing supply is a risky business. Pro- tended not to coincide (Ball and Wood, duction and land purchase decisions have to 1999). In some industries, such as automo- be made on the basis of forecasts of highly biles, international differences in macroeco- uncertain prospects. Secondly, it is also ex- nomic cycles have encouraged firms to tremely difficult to forecast how long market globalise. This has not occurred in house- upswings last and the extent to which they building; presumably, because low scale are dampened or exacerbated by business economies and the importance of local infor- cycle effects. Who, for example, would have

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 902 MICHAEL BALL

Figure 2. Real house price changes in France, the UK and the US, 1985–2000. Sources: France: Ministe`re de l’E´ quipement; UK: Council of Mortgage Lenders; USA: Fedstats.

predicted in the mid 1990s that housing mar- industrial restructuring during downswings kets in many countries in the early 2000s (Maddison, 1992). would help to offset recessionary tendencies in other economic sectors? Thirdly, extensive econometric research shows that price behav- Functional Organisation of the House- iour in the housing market may be irrational building Process in the sense that forecasts of future market There are three prime functional aspects of performance are based on recent experi- housing development: ence—i.e. adaptive expectations (Dipasquale and Wheaton, 1994; Muellbauer and Mur- (1) Residential land development. Land has phy, 1997; Meen, 2001). This may help to to be acquired, the appropriate regulatory explain why the ending of the upswing permissions sought and the site prepared phases of market cycles tends to come as a with infrastructure so that homes can be shock event, which radically alters sub- built on individual serviced plots. With sequent behaviour and housebuilder perform- redevelopment sites, this may require the ance. demolition of an existing structure or The consequences for homebuilders of this wholesale land reclamation. type of market structure are significant. Most (2) Housing production. This entails the ac- importantly, substantial market risk has to be tual building of dwellings—from sub- taken account of in the organisation of pro- structure to completed and fitted-out duction. Furthermore, housing market cycles superstructure. can be the opportunity for spectacular profit- (3) House marketing and sales. Completed making, but equally spectacular loss-making housing is transferred to ultimate users, during downswings. This gives the industry either through sales to individual home- some of the rhythms that Schumpeter sug- owners or via some type of landlord. gested occur in industries over the course of This process involves sales in various the business cycle: organic expansion and types of owner-occupied market (for ex- optimism during upswings and substantial ample, speculative, pre-sales or one-off

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 903

Figure 3. Ovelapping functions in housing development.

contracts with owner ‘builders’) or in to social landlords. In both the large social rental markets (e.g. sales to private land- and private landlord cases, development lords of single or multifamily units or companies could also approach landlords contracts with social housing agencies). with projects they wish to build on land they Sales, thus, can be made either before or have already assembled, or offer off-the-peg after houses are actually built, as con- standardised building designs completed at a tracts or ‘speculative’ sales respectively. fixed price for construction on sites already prepared by the landlord. Figure 3 illustrates these three functions and There are also differences between the or- their interrelationships. Next to each function ganisation of the overall housing develop- in the diagram is a box highlighting the ment process in relation to single-family and relations to other agencies and markets. A multifamily housing. Multifamily structures Venn diagram is used in Figure 3, rather than are generally more complex and capital-in- a sequential series of boxes, to highlight the tensive than single-family ones, because of fact that the three housing development func- the greater need for load-bearing walls plus tions are interrelated and may be institution- common access and other facilities. Multi- ally combined in different ways. family structures are also more likely to be Speculative housebuilding of single-family heterogeneous single or grouped structures dwellings for home-owners clearly demar- on greenfield or redevelopment sites, rather cates the process sequentially from land as- than the simple repetitive built forms com- sembly through to sales. Even there, though, mon with single-family dwelling, suburban homebuilders are keen on pre-sales before expansion. This tends to mean that the devel- starting to build. In social housebuilding, opers of multifamily dwellings have to be land assembly, project conception and larger to gain access to capital. They also finance have traditionally been initiated by have to be more focused on design and con- social housing agencies themselves. In which struction matters than in the single-family case, land assembly and house ‘sales’ take case, where repeat designs and simple con- place at the time of project approval and struction formats are more common. actual building subsequently takes place Such descriptions are inevitably generali- when contracts are let to private building sations, because the possible combinations of contractors. Large-scale private landlords the three functions and their relationships to might adopt a similar procurement strategy other interlinked markets are manifold and

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 904 MICHAEL BALL there is little point in describing them all in functional distinction between land develop- detail here. Several features, nonetheless, are ment and actual housebuilding, because both common in descriptions of them require particular types of management and finance. One of the most important factors —a network of interrelationships is de- relating to production is the employment re- scribed; lationship. —the physical characteristics of the overall Housebuilding involves a construction housing development process are process and so requires a specific range of significant influences on housebuilding or- construction-related labour competences. ganisational relationships; Building involves a planned sequence of ac- —economic and financial factors play an tivities by a workforce of varying skill levels, important role in determining organisa- that utilise specific pieces of equipment. As a tional structures; and production activity, it generates significant —public policy with regard to housing information asymmetries between those tenures, land markets and building pro- undertaking building work and those manag- cesses have significant effects on institu- ing them. It is expensive to monitor individ- tional housebuilding arrangements. ual work activities directly on a The overriding point is to show that different housebuilding site and machines do not set ways of organising the housing development the pace, so that managers do not have full process are associated with distinct housing control over work effort and quality. outcomes. It is consequently dangerous to However, most tasks in housebuilding assume that each type of housebuilding ex- are amenable to being broken down into hibits similar behavioural characteristics and specific, priced jobs and can be let out to necessarily falls under the rubric of a com- sub-contractors and other specialists (Ball, petitive market. One way of exploring these 1996). varying approaches to housebuilding is to Sub-contracting resolves several infor- categorise each of them as an institutional mation asymmetry problems, because man- structure of housing provision. An institu- agers only have to check that work has been tional structure of housing provision specifies done properly and on time and, then, pay out the organisations associated with a way of the agreed sum, rather than have to monitor creating housing and the rules and con- work effort continuously. Sub-contracting straints influencing behavioural relations be- encourages standardisation of work tasks and tween them. In this way, the functions of simple, easily repeatable, dwelling designs, agencies in the production process can be so that tasks can be broken down into uni- mapped out for each type of housing devel- form sub-contracts and assessed more easily. opment (Ball, 1998). This has the benefit of Housebuilders can impose financial penalties recognising that any agency or firm in hous- on non-performing sub-contractors and build ing development is involved in a network of up long-term relationships with core sub- interrelationships and enables a mapping-out contractors to ensure that they self-police of the constraints and influences on individ- their quality and prices. These sorts of sub- ual agency behaviour. Specific theories are contracting relationship, and their benefits, still needed, none the less, to understand have become famous since the discovery of what institutional arrangements are important ‘Japanese’ manufacturing methods in North and why they exist. America and Europe in the 1980s (Ricketts, 1994; Kay, 1996). Yet, similar ones have existed since, at least, the 19th century in Housing Production Influences on Institu- many countries’ housebuilding industries tional Structures of Housing Provision (Price, 1980). Sub-contractors themselves of- A useful starting-point when examining ten hire part of their workforces on time, housebuilding institutional structures is the price and task-specific bases—whilst, at the

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 905 same time, employing a core of competent evolved information networks about job skilled workers and supervisors. So, a hier- opportunities which facilitate resource archical tree of work patterns—embodying transfer. These are often informal in nature. market and non-market relationships—can Overall, consequently, housebuilding in- frequently be observed in housing production. puts, including labour, should be more Another common employment option for fully and effectively employed in countries housebuilders is to hire workers on a tempo- that have construction industries with rary hourly or daily basis. In this situation, the ‘flexible’ labour practices. In contrast, problem of directly monitoring work effort countries that require direct employment arises again. This again encourages the use of through labour laws are likely to impose standardisation and simple techniques of pro- significant periods of idle time on workers. duction with easily observed results, so that This idle time may occur through more relatively unskilled workers can be hired— extensive periods of unemployment. Firms, and fired if they do not keep up with normal for example, may be reluctant to employ work rates. Sub-contracting and day-working extra workers because of future workload considerably aid the flexibility of the house- uncertainties or due to the implicit cost building industry in coping with uncertain associated with hiring when subsequent demand. When housebuilders’ output falls, firing is expensive to undertake. Alterna- they do not have to lay-off workers, but tively, the idle time may lower firm instead merely issue fewer contracts or hire efficiency, with firms holding onto under- fewer workers; a similar situation exists with employed staff because it is too expensive to plant when it is hired. lay them off. Such self and temporary hiring employ- Unfortunately, there are no rigorous, em- ment techniques have implications for the pirical cross-country studies of employment types of production method used, as the latter outcomes in housebuilding and construction have to be effective in the context of those in general by which to test these labour employment practices. Generally, this means market hypotheses. However, both theory and that techniques are relatively simple, do not the fact that building firms adopt the flexible use much sophisticated capital equipment and model whenever they can suggest that the change slowly. When equipment is complex enforced direct employment option is a more or expensive, it is hired complete with skilled expensive and less efficient mode of oper- operatives to avoid any potential skill ation. Furthermore, direct employment labour difficulties. The same occurs with more tech- laws impose considerable costs on those nically complex materials or fittings. workers that cannot find jobs or try to find The capital tied up in housing production, work again when they are laid off. therefore, is generally small for employment In countries and regions where the climate practice reasons, as well as to ensure makes it impossible to build in the deepest flexibility in face of varying demand. Surveys winter months, for example, the large-scale show that production capital is often financed switching of resources between housebuild- out of retained profits and short-term credit ing and other activities is obviously impera- loans, although land stocks are usually too tive at different seasons of the year. expensive to be funded in this way (Hille- Traditional migrations are made between, brandt, 1971). say, summer housebuilding and winter for- Such an organisation of production estry or between educational establishments maximises the possibility of resources being and summer vacation building sites. Simi- fully employed elsewhere when one parti- larly, in countries with large agricultural sec- cular housebuilder has insufficient work. tors that have heavy demands for labour only Furthermore, workers and plant exist in an at peak times, such as harvests, another pat- institutional framework of well-developed tern of seasonal job shifts in and out of sub-contract markets—a framework that has housebuilding tends to occur.

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 906 MICHAEL BALL

Some have argued that labour sub- onomies. Firms’ strategic choices over mar- contracting is bad for building quality and ket segments in which to be active, therefore, productivity, because it leads, first, to rushed simultaneously have implications for their work and, secondly, to a slow decline in the internal organisational structures and man- skill level of the construction workforce agement needs. Yet, as scale economies are (Clarke and Wall, 1996). The first claim is low, the benefits of the extra size that may suspect because of the information asymme- result from winning work across a variety of try problem, mentioned earlier. The second types of construction is limited or may even argument is a ‘tragedy of the commons’ ex- be negative, as management diseconomies ternality type argument in which no-one is set in. When firms diversify, consequently, prepared to train skilled workers because possible increased unit costs in production they will always be poached by others. It is may have to be compensated for by other difficult to see how the nature of the employ- benefits, such as risk-spreading. ment relation changes that argument much, Sub-contracting, it can be concluded, has however, because the incentives for a trained significant impacts on firm organisational worker to move to a better-paid job would structures. The practice considerably en- not alter as long as firms free-ride on other’s hances the ease of entry and exit from house- training expenses.1 building and, hence, the degree of The sub-contracting of tasks has important competition within the industry. Yet, some implications for the firm structure of house- countries, especially in Europe, ban exten- builders. It means that, at low levels of out- sive sub-contracting and their employment put, housebuilders can be very small entities laws require that workers are hired on a indeed. A one-proprietor entity, for example, permanent basis. This is the case in France, can hire outside help for most construction for example, although some sub-contracting tasks. Even larger firms, when measured by has been permissible there since 2001. The their output, need only have a limited direct arguments above would suggest that such a production workforce, consisting primarily ban leads to higher construction costs. Of of those managing sub-contractors or ensur- particular interest in the context of the argu- ing a steady flow of materials or skilled ments being made here is that the legislation operatives fixing incomplete or poorly done would also be predicted to create a different tasks.2 array of types of housebuilding firm from Many firms find that they can also keep those prevalent in countries with extensive management costs down by specialising in sub-contracting or day labour.3 only one form of housebuilding (such as In the French situation, it is to be expected single-family owner-occupation or large that most firms obey the spirit of the law and multifamily structures). This arises because employ large, direct workforces. On the basis different types of built structure often require of the arguments outlined above, therefore, distinct construction processes and, hence, they would be predicted to be bigger in terms of skills to manage them. When structures are employment size than in the sub-contracting more complex, moreover, this complicates institutional framework and face higher design, materials procurement and logistics. labour and plant-related costs. Such cost pres- Building firms frequently outsource pro- sures, furthermore, encourage firms to diver- fessional tasks, such as design and engineer- sify in order to steady workloads and cover ing work. overhead costs. They pass the resultant extra Some firms, nevertheless, straddle differ- costs onto clients in higher building prices. ent types of building. When they do so, This argument, consequently, predicts that construction firms frequently utilise the M- the scale and scope of housebuilding firms form type of management structure will be greater in countries with strict em- (Williamson, 1981), with quasi-independent ployment legislation and housebuilding costs subsidiaries to limit management disec- in such countries will be relatively high.

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 907

Some French evidence seems to confirm this understandably they regard the employing of prediction, particularly the fact that large cheaper, illegal, casual labour as unfair com- construction firms have increasingly moved petition. It claims that over half of single- into housebuilding (Campagnac, 1992; Cam- family dwellings are constructed illegally pagnac and Winch, 1995) as part of a di- (Baˆtirama, 2002). Thus, labour legislation, versified portfolio of construction activities. ostensibly aimed at improving the efficiency Some housebuilders in restrictive labour of the construction industry, actually miti- legislation countries, however, choose to ig- gates against competition based on relative nore such laws and illegally sub-contract. efficiency and, instead, encourages it on the This leads to ‘formal’ and ‘informal’ con- basis of whether or not to conform to the struction sectors. Informality is prevalent in law. housebuilding because of the frequently Traditional small builders and small pro- small scale of activity, which makes it prietor-run firms tend not to be affected by difficult for governments to police. This du- restrictive labour legislation. Consequently, alism can be seen, for example, in Germany, they are protected to a degree from compe- particularly between those building workers tition from larger, more ‘modern’ developers permanently living in the country and the and builders. This may help to explain why a large influx of extra building workers from traditional type of housebuilding, owner de- lower-wage European countries. In France, velopment, is so prevalent in countries like the problems of monitoring informal building France and Germany. With owner develop- labour have helped to generate further legis- ment, individual home-owners buy plots of lation. Since 1990, for example, the signing, land and then hire some organisation to build and registering, of a ‘construction contract’ is their home for them. It may be a builder, who compulsory before a house is built. The provides an off-the-peg design or an architect claimed objective is to provide better protec- and a group of independent ‘jobbing’ tion for clients, but it also ‘formalises’ build- builders specialising in particular house- ing work, so that it has to conform to building trades. Owner development repre- employment law. The system, in practice, sents around half of all the new dwellings has unsurprisingly proved difficult to admin- produced each year in France. In respect of ister. The professional body representing single-family housing, the data suggest that housebuilders (Fe´de´ration Franc¸aise des most in France are owner-developed (Table Constructeurs de Maisons Individuelles) has 2). In Germany, many owner-occupied recently produced a statement on the need to houses are built in a similar way and it is ensure full enforcement of the law, because common in Europe’s Mediterranean shore-

Table 2. Types of new development and dwelling in France

Percentage share Type of Starts in 2001 Percentage share Type of development in 1999 dwelling (000s) in 2001

Owner-developera 52 Single houses 172 48 Construction firms 27 Grouped houses 43 12 Social and semi- 12 Flats 128 36 public developers Other 9 Other 14 4

Total 100 357 100

aOwner-occupiers that commission a dwelling to be built on a land plot they already own. Sources: Dwellings, INSEE; Developments, Ministe`re de l’e´quipement, des transports et du logement, Paris — DAEI-SES: Sitadel.

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 908 MICHAEL BALL line countries. This owner self-build market through regulatory procedures and the site is ideal terrain for the small builder, but they has to undergo any necessary and expensive share it with larger firms that sell and build works required to enable individual building standardised designs, sometimes in a par- plots to be prepared. tially prefabricated form. The construction and project-risk charac- Even in countries with a much higher pre- teristics of individual sites can be pooled by ponderance of mass-market housebuilders, firms by developing more than one site at a there are some individuals who buy land and time. Greater firm size may also lower bor- hire agents to design and build their homes rowing costs as lenders can more easily (Barlow et al., 2001). In those countries, monitor loans. The land development pro- however, it is probable that such methods cess, consequently, suggests that greater firm tend to occur at the upper end of the market; size is a benefit. The size of the risk-pooling yet, as has been seen, it is the owner-occu- benefits depends on the spatial level at which pied mass market in France. risk-pooling ceases to outweigh any counter- acting management diseconomies of scale. Most market information suggests that this point is probably reached by, at least, the The Land Market and Residential Devel- regional level. Any higher spatial pooling opment advantages depend on the degree of corre- The land market is like any other in that lation between regions’ housing markets. In a prices are determined by the interaction of country like the UK, the literature suggests demand and supply. In the aggregate, the that regional housing markets move quite land market can consequently be represented closely together, although house prices are in the normal competitive way. What is im- more volatile in the South East which leads portant for examining the organisational rela- other regions over the course of the housing tionship between the land market and cycle (Meen, 2001). Builders may conse- housing development, however, is the way in quently be able to pool risks by operating in which actual land plots are bought and sold. the South East and several other regions A distinctive feature of the land market is simultaneously. In Australia and the US, the that each building plot has unique locational regional pattern is more variable (Pol- characteristics. Land sites, consequently, are lakowski and Ray, 1997; Meen, 2001), so imperfect substitutes for each other. Partly again there may be risk-pooling benefits of because of this locational characteristic, what housebuilders adopting a multiregional strat- is built on any particular site is not a standard egy. Multiregional strategies, none the less, ‘can of beans’ style product, but a unique set do not imply the need for ‘national’ pres- of built structures. Developers consequently ences by operating throughout a country. In have to conceive of a project and make the this respect, housing is unlike many other appropriate calculations about likely costs, consumer goods industries, where the im- uncertainties and returns. They also have to portance of brands and production technolo- take into account the effects of any new gies encourage operation at the national and project on the firm’s overall strategic objec- international levels. tives. Land development outcomes are highly Land-owners and developers enter negoti- risky. An indication of the greater riskiness ations about the price and timing of land of land development can be seen in the gen- purchase. Land acquisition can either take erally higher volatility of residential land the form of outright purchase or an option, prices compared with house prices. which is a futures contract to acquire land at The land development process is often a later date subject to specified conditions. more capital-intensive than housebuilding, The nature of the land purchase negotiation which intensifies the risks. Land has to be has implications for firm organisation. In purchased, time has to be spent passing those negotiations, key issues are the trans-

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 909

Figure 4. Negotiating over development profit and land price. parency of a site’s development value and Land banks may also offer developers the current phase of the housing market cy- negotiating advantages, because they mean cle. If there have been many similar recent that a developer is not forced to buy but can land transactions in a locality, a land’s value instead temporarily run down land stocks. as a residential site is easy to identify for Land banking helps consequently to facilitate both parties, so that the negotiations should a strategic view by developers of the land be straightforward. In many situations, how- market. One component of strategic land ever, there is greater uncertainty over land holdings may be spatial land strategies. development value. In such situations, nego- These put developers in a good negotiating tiation is more significant and strategic. position with land-owners, because the latter Figure 4 illustrates some factors that will know that the former can take their influence land-owner and developer nego- housebuilding activity elsewhere. They also tiation strategies. Land-owners’ positions are form part of interdeveloper competition—a influenced by the state of the local land large development in a locality by one devel- market and their alternative options, includ- oper, for instance, may block a similar move ing the possibility of holding onto the land by another one for fear of creating local until better offers come along. Developers, in oversupply. contrast, in order to strike a good bargain are Many of the developer negotiation factors likely to emphasise the potential riskiness of just highlighted suggest that there are their proposed development. They may also advantages to greater organisation size in gain through their superior knowledge of the residential land purchase and development. local housing market and the planning and This is because larger enterprises have em- other regulatory systems, which enable them ployee skill-bases, capital-bases and land to maximise both their and land-owners’ re- banks that enable them to spread risks, lower turns if land-owners co-operate. Impact fees financing costs, improve negotiating posi- and planning obligations payable to the local tions with land-owners and facilitate strategic municipality further complicate land negotia- actions. This does not mean that smaller tions. With these, developers again generally developers are necessarily competed out of have superior knowledge to land-owners and the land market, rather that their operations they can assist land-owners over such mat- are likely to be more risky and credit- ters, if the price is right. constrained.

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 910 MICHAEL BALL

Land Development and Housing Construc- or semi-public authorities key land develop- tion ment agencies for all types of property devel- opment: for example, in the Netherlands, An obvious producer firm distinction in France, Germany and Sweden (Barlow and housing is between those enterprises that King, 1992; Golland and Boelhouwer, 2002). purely undertake land development, those These bodies, generally under municipal that solely build housing by acquiring pre- control, took over the function of traditional serviced building plots and those enterprises private-sector land developers, relying for that combine both the land development and funding on central or regional government building functions. In some countries, the grants and public-sector banks for loans. In actual division between these types of firm, most of these countries, however, such land and the relative market shares they have, is development practices generally did not be- strongly influenced by government policies. come universal, so that purely private land Yet, in most situations, market forces seem developers continued to exist in large num- to be the prime cause of the separation or bers. combination of land development and house- building. This section provides some sugges- tions as to what those forces might be. Private Land Developers and Builders: Separation and Combination Purely private-sector land developers are, of Regulation. Government policy has the course, the most common type of developer greatest influence in Europe. This is not sur- in market economies, apart from in a handful prising as, in many European countries, land- of European countries. Yet, it is still possible use planning and housing policies are to observe significant differences in the insti- strongly interventionist. The introduction of tutional structure of the development pro- large-scale social housing programmes in the cess. Most notably, there is often an middle decades of the 20th century, for ex- organisational separation between land de- ample, was, in part, based on criticism of velopment and housing construction; at the pre-existing housing development and its de- same time, the two aspects are also often sign, as well as constituting a sustained at- undertaken by one enterprise in a combined tempt to provide better housing for low- way. These two forms, independent develop- income groups (Swenarton, 1981). As a re- ers and builders and combined developer- sult, social housing landlords became major builders may even exist in the same housing land developers from the 1920s onwards, as market. The separate arrangement is com- well as large-scale landlords. They undertook mon, for example, in Australia and the US,4 the development function in new housing, for and it was prevalent in Britain’s primarily many years priding themselves on their rev- private rental housing system prior to 1914 olutionary estate designs and built forms; (Ball and Sunderland, 2001). In contrast, in whilst they overwhelmingly hired private modern Britain, development and building building contractors to build the actual struc- are now generally combined under the con- tures. There was some attempt to munici- trol of one housebuilding firm that buys sites, palise housebuilding completely, by bringing erects homes and sells them. Why should it under the control of state housing agencies such distinctions occur and why are they or local authorities. Such moves tended to important? have only limited impact in new building, The relative riskiness of the two aspects although they were far more influential in of development and building and the com- social housing repair and maintenance petitiveness of land markets in different (Briscoe, 1990). countries may together account for such In the 20th century, furthermore, policies organisational differences. It was noted were introduced in Europe that made public earlier that land development was a more

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 911 risky activity than that of building, so returns by several studies of housebuilding in the US should be larger for it to compensate for the and Canada, where such characteristics tend extra risk. So, if entry and exit is easy in to prevail (Clawson, 1971; Grebler, 1973; these two aspects of housebuilding, competi- Buzzelli, 2001). A recent study, furthermore, tive pressures should lead to a simple div- suggests that there are systematic spatial ision of land development and building into variations in the average size of home- different types of enterprise, because this is builders in such markets; these depend on, the least-cost option in a risky market. for example, whether the urban area is grow- The argument is as follows. When rela- ing, when firms tend to be larger, or contains tively small housebuilders can buy serviced small-scale, development restrictive, local sites from land developers, they can enter governments, where firms tend to be smaller and exit the industry easily without the en- (Somerville, 1999). cumbrance of land development procedures When land supply is restricted, in contrast, and associated capital costs. They can choose as it is in most of urban Britain, the attrac- their sites if they wish from more than one tiveness of combining housing development developer, spreading project-specific risk and and building is greater. In this market con- encouraging competition between land de- text, holders of residential land sites develop velopers. In competitive land and housing a degree of monopoly power through the markets, with, say, a ready supply of land they own. With restricted land supply, greenfield suburban land, project-specific the riskiness of development is diminished as risk may well be high because of the compe- individual project risk is less than in the tition from other local developers. Builders’ competitive market case, and there is less and developers’ decisions to build and put chance of spatial competition between devel- land plots onto the market, when added alto- opers and builders generating overbuilding gether, might also lead to excess supply, so However, there is still much risk associated that market-based risk is similarly high. Be- with general market volatility. ing able to sell plots to builders who then So why combine development and build- build houses and sell them, spreads the risk ing, when this monopoly power could be of losses between developers and builders. exercised by a pure land developer? The Each has to commit less capital for their answer could be that combining development respective stages of the housebuilding pro- and housebuilding heightens firms’ cess and, equally importantly, the time that influences on local housing and land markets. capital has to be committed before a profit If, for instance, developers sold plots on to can be taken is far less than for firms that separate housebuilders whenever they were both develop and build. offered the current market price for land, The suggested hypothesis here, in sum- then the latter would determine when and mary, is that in competitive markets differ- how much new housing was supplied. In ences in risk and returns between the contrast, unified developer/housebuilders can development and building aspects of house- command significant residential market building create a tendency for the land devel- shares in localities with limited land supply opment and construction sides of and, so, have a greater chance of influencing housebuilding to be separated. Combined de- both housing output levels and local land veloper-builders may well still exist in com- prices, because conditions of monopolistic petition with such firms, but the predominant competition would then prevail in both mar- tendency would be towards separation. In kets. such market contexts, furthermore, there should be a large turnover of housebuilding Owner-occupied Housebuilding in Britain and real estate development firms, because of the ease of entry and exit and the scale of In Britain, land supply is frequently highly risk. Such a firm structure has been reported restricted, so the market-position benefits of

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 912 MICHAEL BALL combining development and building are The Dynamics of Industrial Change brought to the fore and large regional house- builders predominate in new housing devel- The ease of entry and exit from house- opment. Such developer-builders frequently building and the riskiness of land develop- build up large strategic land banks and, be- ment mean that housebuilding is an industry cause of their size, have a significant with substantial enterprise turnover, as noted influence in the residential land market itself. above. Such churning leaves few firms un- Nevertheless, as entry and exit from both touched. Hardly any of the major British housebuilding and land development are still housebuilders have had a continuous exist- quite easy, the oligopolistic position that ence in approximately the same form for the some UK housebuilders manage to create past two decades. probably does not lead to large monopoly Industrial change is driven by the property profits for them, but neither does it force market cycle and regulatory changes, rather firms to operate in the most efficient ways than by the technological factors, scale and possible as competition is blunted. scope economies and international trade Thus, an ability to influence strategically pressures that are common in many manufac- the markets for residential land and new turing industries (Chandler, 1977; Porter, housing is a likely cause of the combination 1990). During upswings in the housing mar- of the developer and builder roles in modern ket, particularly if they are long-lasting, new British housebuilding. British planning au- entrants flood into housebuilding and exist- thorities, in any case, are said to have a ing firms expand. This often occurs alongside preference for dealing with larger, profes- hype that the industry has changed, or that sionally staffed, building firms, because they some firms within it have discovered a new are easier to deal with, more knowledgeable formula, or that for some reason size has now of planning practices and can be relied on to become more important. The subsequent conform to planning requirements (Ball, downturn, however, leads to a shrinkage in 1983; Rydin, 1985). both the number of active firms and the size The rank–size distribution of house- of those remaining, with the unluckiest or builders in mid 1990s Britain shows that it is most overcommitted and overoptimistic ones highly skewed with a number of large pro- failing. ducers and a tail of smaller, but still The land market plays a part in generating significant, firms—most of which are com- firm restructuring, especially when land is in bined developer-builders (Ball, 1996). Sug- temporary or long-term short supply. During gested reasons for scale economies for such housing market upswings, firms may find enterprises were considered earlier and relate that it is cheaper and quicker to acquire land to risk-pooling, purchasing and finance econ- by taking over another builder and its land omies, and strategic behaviour. There are bank than through direct purchase in the land also size benefits through the marketing market. Similarly, surviving firms may ac- gains of having a well-advertised sales port- quire the land and other assets of the failing folio of new housing that is ‘brand imaged’. firms cheaply during housing market slumps Most UK housebuilders are regionally based (Ball, 1988). with a complete national presence being rare, Occasionally, changing demands for hous- as can be seen in Table 3. Smaller British ing create new market segments in which housebuilders, of which there are thousands, specialist firms can grow up. The inner-city tend to survive in this market environment ‘loft’ and ‘luxury apartment’ markets have through the idiosyncrasies of their owners, already been mentioned. Retirement housing by taking greater risks or through specialis- is another case where specialists have ing in small sites where larger firms are at emerged with rising retiree wealth and the either a cost, information or quality disad- growing need for specialist care facilities vantage. with ageing populations. Both these markets,

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 913 Persimmon Wimpey David Wilson Haslam BarrattWimpey Barratt Taylor Woodrow Redrow Bellway Miller David Wilson Taylor Woodrow Redrow South West West Midlands Yorkshire Bellway Westbury Redrow Persimmon Westbury Persimmon Westbury Prowting Persimmon Barratt David Wilson Wimpey Anwyl McCarthy & Stone Watkin Jones Harron Wimpey Barratt Taylor Woodrow Barratt Bellway Wilcon Persimmon Bovis Wilcon Morrison McCarthy & Stone Cannon Kirk Crest Nicholson Crest Nicholson Miller Persimmon Berkeley Scotland South East Wales Barratt Redrow Taylor Woodrow Rialto Stewart Milne Bovis Miller Cala C M Yuill Dunelm Sir Robert McAlpine Bett Wimpey Taylor Woodrow Wimpey David Wilson Wimpey J S Bloor Rank of top 10 housebuilders, by British region, 2001 (largest first) Bellway Persimmon Barratt Persimmon Bowey Taylor Woodrow Bellway Wilcon Berkeley Barratt Table 3. David Wilson Redrow Wilcon East Midlands North West North East Persimmon Miller Westleigh (2002). Taylor Woodrow Henry Boot Centex Excludes housing associations. Based on detailed planning permission submissions. Persimmon Barratt E. Anglia London WimpeyWilconBovis Bellway Berkeley Crest Nicholson Bellway Wimpey Barratt AbbeyBellway Laing Wimpey J S Bloor David WilsonJS Bloor Weston Centex BennettKier Fairview McCarthy & Stone David Wilson Broseley Notes: Source: Housebuilder

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 914 MICHAEL BALL moreover, have benefited from regulatory the range of functions undertaken by individ- change, particularly in planning and infra- ual enterprises and the size of firms. structure provision. Many conclusions must remain specula- Regulation in the broadest sense, further- tive, however, because there is insufficient more, is not simply a government domain, empirical evidence to test hypotheses rigor- but can arise, for example, from changing ously and little pre-existing literature. How- financial market practices. From the 1950s to ever, the arguments made here correspond to the 1990s, many large UK housebuilders general observations about the structure of were parts of far-larger construction con- countries’ housebuilding industries. glomerates. This is rarely the case now. This Overall, whilst it may not be possible sim- development has probably been aided by the ply to treat housebuilding as a competitive demand for greater financial transparency industry with the standard structure– from construction firms by modern banking conduct–performance paradigm, never- and capital market lenders. Transparency is theless, standard economic analysis is still hard to make credible in a multifunction useful in explaining observed institutional construction enterprise, which must have en- structures. This, of course, does not mean couraged the shift of firms towards clearer that housing supply is necessarily riddled core functional activities (Ball, forthcoming). with monopoly, rather that it is quite com- plex and needs detailed empirical investiga- tion before robust conclusions can be Conclusions reached about its operation and performance. This paper has examined reasons for differ- Differences in institutional structures af- ences in the organisational structure of fect supply responses and general levels of housebuilding in different countries. It fo- industrial efficiency. Not too much can be cused analysis on potential economies of said on a comparative basis about relative scale, market factors, information asym- supply responses internationally, apart from metries, regulation and risk. The great var- the fact that they are likely to be highly iety of ways in which housing is built, it was variable and depend on the competitiveness argued, could be explained in terms of a of the land market. Furthermore, public pol- limited range of concepts commonly used in icy towards housing and land markets tends industrial economics. Market instability is an to neglect any consequences for housing sup- important factor, but is insufficient to provide ply. Interventions by governments into a full explanation because it is a common housebuilding, nevertheless, have not had feature of durable goods industries, most of much success when tried. which are not organised in the same ways as housebuilding. Locational specificity is im- Notes portant, especially with regard to the land 1. Compulsory industry-wide training levies market. The type of housing being built and and grants to firms offering training and the markets in which they are sold are further individuals undertaking it are policies aimed influences. Information is particularly im- at internalising these externalities—as, for portant in explaining the economic character- example, with the Construction Industry istics of housebuilding, because of the nature Training Board in the UK (www.citb.org. uk). of both the development and production pro- 2. Sub-contracting may also arise as a means of cesses. Strategic behaviour enters the equa- avoiding taxes in contexts where the effec- tion, particularly through behaviour with tive tax rate on the self-employed is lower regard to the land market and residential than on directly employed workers. Such tax development strategies. Regulation in labour effects, for instance, contributed to the rise of labour-only self-employment in the UK from markets also has considerable consequences the 1970s onwards. Conversely, tightening of for firm structures. Land availability and the the tax rules in recent years and legislation impact of planning regimes on it, both affect creating a more flexible labour market have

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 MARKETS AND THE HOUSEBUILDING INDUSTRY 915

led to a growth again in the share of direct in the Twenty-first Century. York: Joseph employment. These tax considerations, how- Rowntree Foundation. ever, do not negate the points made here BAˆ TIRAMA (2002) La plupart des maisons sont about sub-contracting, because directly em- construites dans l’ille´galite´, (http://www. ployed workers obviously can be employed batirama.com/data/12042002/12042002–161715. by sub-contractors as well as by ‘main con- html). tractor’ housebuilders. BRISCOE,G.(1990) The Economics of the Con- 3. The information on France is based on dis- struction Industry. London: Mitchell Publish- cussions with officials at the French Ministry ing. of Housing and the French Federation of BUZZELLI,M.(2001) Firm size structure in North Developers (FNPC) in 2002, as well as web American housebuilding: persistent deconcen- searches (Batirama, FNAIM). The analysis tration, 1945–98, Environment and Planning A, and opinions expressed here, however, are 33(3), pp. 533–550. the responsibility of the author alone. CAMPAGNAC,E.(1992) Les cadres dans la con- 4. An interesting history of such a developer, struction en France: politiques d’entreprise et Newland Communities, can be found at e´volution des professions, in: M. BALL,E. www.newlandcommunities.com. CAMPAGNAC and G. GIALLOCOSTA (Eds) E´ vol- ution des professions et politiques d’emploi des cadres dans les entrerprises de baˆtiment en Europe (France, Grande-Bretagne, Italie), References pp. 11–119. Paris: Ministe`re de l’e´quipement, du logement et des transports, plan construction BALL,M.(1983) Housing Policy and Economic et architecture. Power: The Political Economy of Owner Occu- CAMPAGNAC,E.and WINCH,G.(1995) The orga- pation. London: Methuen. nization of building projects: an Anglo/French BALL,M.(1988) Rebuilding Construction: Econ- comparison, Construction Management and omic Change and the British Construction Economics, 13, pp. 3–14. Industry. London: Routledge. CHANDLER,A.D.(1977) The Visible Hand: The BALL,M.(1996) Housing and Construction: A Managerial Revolution in American Business. Troubled Relationship. Bristol: Policy Press. Cambridge, MA: Harvard University Press. BALL,M.(1998) Institutions in British property CHIANG, Y.-T. and TANG, B.-S. (2002) ‘Sub- research: a review, Urban Studies, 35(9), pp. 1501–1518. marines don’t leak, Why do buildings?’ Build- ing quality, technological impediment and BALL,M.(forthcoming) Markets and Institutions in Real Estate and Construction. Oxford: organisation of the building industry in Hong Blackwell. Kong, Habitat International, pp. 1–17. CLARKE,L.and WALL,C.(1996) Skills and the BALL,M.and MORRISON,T.(2000) Housing in- vestment fluctuations: an international compari- Construction Process: a Comparative Study of son, Housing, Theory and Society, 17(1), Vocational Training and Quality in Social pp. 3–13. Housebuilding. Bristol: Policy Press. BALL,M.and SUNDERLAND,D.(2001) An Econ- CLAWSON,M.(1971) Suburban Land Conversion omic History of London, 1800–1914. London: in the United States: An Economic and Govern- Routledge. mental Process. Baltimore, MD: Published for BALL,M.and WOOD,A.(1999) Housing invest- Resources for the Future by the Johns Hopkins ment: long run international trends and volatil- University Press. ity, Housing Studies, 14, pp. 185–211. DIPASQUALE,D.and WHEATON,W.(1994) Hous- BALL, M., MORRISON,T.and WOOD,A.(1996) ing market dynamics and the future of house Structures, investment and economic growth: a prices,. Journal of Urban Economics, 42, long-run international comparison, Urban Stud- pp. 1–17. ies, 33, pp. 1687–1706. DONNISON,D.(1967) The Government of Hous- BARLOW,J.and KING,A.(1992) The state, the ing. Harmondsworth: Penguin Books. market and competitive strategy: the house- GANN,D.(1996) Construction as a manufac- building industry in Britain, France and turing process? Similarities and differences Sweden, Environment and Planning A, 24, between industrialised housing and car pro- pp. 381–400. duction in Japan, Construction Management BARLOW,J.G.and DUNCAN,S.(1994) Success and Economics, 14, pp. 437–450. and Failure in Housing Provision: European GLENDINNING,M.and MUTHESIUS,S.(1994) Systems Compared. Oxford: Pergamon Press. Tower Block: Modern Public Housing in BARLOW,J.G., JACKSON,R.ET AL. (2001) Homes England, Scotland, Wales and Northern to DIY for: the UK’s Self-build Housing Market Ireland. London: Yale University Press.

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015 916 MICHAEL BALL

GOLLAND,A.and BOELHOUWER,P.(2002) Specu- and busts in the UK housing market, Economic lative housing supply, land and housing mar- Journal, 107, pp. 1701–1727. kets: a comparison, Journal of Property POLLAKOWSKI,H.O.and RAY,T.S.(1997) Hous- research, 19(3), pp. 231–251. ing price diffusion patterns at different aggre- GREBLER,L.(1973) Large Scale Housing and gation levels: an examination of housing Real Estate Firms; Analysis of a New Business market efficiency, Journal of Housing Enterprise. New York: Praeger. Research, 8(1), pp. 107–124. HARLOE,M.(1995) The People’s Home? Oxford: PORTER,M.(Ed.) (1990) Competitive Advantage: Blackwell. Creating and Sustaining Superior Perform- HILLEBRANDT,P.M.(1971) Small Firms in the ance. New York: Free Press. Construction Industry. London: HMSO. PRICE,R.(1980) Masters, Unions and Men: Work JACKSON,A.J.(1991) Semi-detached London: Control in Building and the Rise of Labour, Suburban Development Life and Transport 1830–1914. Cambridge: Cambridge University 1900–39. Didcot: Wild Swan. Press. KAY,J.(1996) The Business of Economics. RICKETTS,M.(1994) The Economics of Business Oxford: Oxford University Press. Enterprise. London: Harvester Wheatsheaf. LE GALE` S,P.(2001) Urban governance and policy ROLFE,S.and LEATHER,P.(1995) Quality networks: on the urban political boundedness Repairs: Improving the Efficiency of the Hous- of policy networks. A French case study, Pub- ing Repair and Maintenance Industry. Bristol: lic Administration, 79(1), pp. 167–184. Policy Press. MADDISON,A.(1992) Dynamic Forces in Capital- RYDIN,Y.(1985) Residential development and ist Development: A Long-run Comparative the planning system: a study of the housing View. Oxford: Oxford University Press. system at the local level, Progress in Planning, MALPEZZI,S.(1990) Urban housing and financial 24, pp. 33–59. markets: some international comparisons, SOMERVILLE,C.T.(1999) The industrial organiza- Urban Studies, 27(6), pp. 971–1022. tion of housing supply: market activity, land MARTIN,S.(1994) Industrial Economics: Econ- supply and the size of homebuilders, Real omic Analysis and Public Policy. Basingstoke: Estate Economics, 27(4), pp. 669–695. MacMillan. SWENARTON,M.(1981) Homes Fit for Heroes: MEEN,G.(2001) Modelling Spatial Housing Mar- The Politics and Architecture of Early State kets: Theory, Analysis and Policy. London: Housing in Britain. London: Heinemann Kluwer Academic. Educational. MOLHO,I.(1997) The Economics of Information: WILLIAMSON,O.E.(1981) The modern corpor- Lying and Cheating in Markets and Organiza- ation, origins, evolution, attributes, Journal of tions. Oxford: Blackwell. Economic Literature, 19(4), pp. 1537–1568. MUELLBAUER,J.and MURPHY,A.(1997) Booms

Downloaded from usj.sagepub.com at UNIV OF GUELPH on April 2, 2015