I.1 THE ARGUMENT
Now in the eighth decade of its existence, the World Bank continues to attract huge levels of attention. Over time the visibility of the Bank has increased substantially, with its press releases now circulating to over 5000 journalists and each of its @WorldBank tweets reaching close on 1 500 000 interested parties.1 Mirroring this high level of public interest, policy makers, representatives of non-governmental organisations, and academics alike all devote substantial resources to trying to understand what the Bank does, why it does it, and how its development impact can be improved.
The bulk of the literature on the Bank that has been generated over the years is heavily critical in nature. Throughout much of its history, the organisation has been in the unenviable position of having vocal critics positioned on both its political left and its political right. For some, the organisation works to reproduce and extend a version of global capitalism that tramples on the needs of the poor. For others, its lending serves to prop up corrupt and ineffective governments – governments whose performance would be better improved through the unmediated application of domestic and international market forces.
Through this book I add to this stock of critical evaluations of the World Bank. I am particularly concerned by the ‘upward drift’ that has taken place in the targeting of the organisation’s lending for housing. In the early decades of its operations in this policy area, the Bank’s projects sought to directly improve the housing conditions of lower-income groups across developing and emerging market economies. The increasingly narrow focus on the extension of mortgage markets, which has become embedded over the last two decades, now serves to target assistance toward the middle- and upper-income groups who are able to access formal networks of housing finance. By and large, lower-income groups have remained excluded from these circuits of capital.
In addition to delivering this critical message on pro-poor targeting in the Bank’s housing engagements, I also aim through the book to advance an analytic contribution that helps us to expand our understanding of the dynamics that shape operational practice in this extremely important institution. I contribute to the body of international organisation-focused scholarship by suggesting that, in the realm of lending for housing, changes in World Bank practice have been driven by a process of ‘constrained experimentalism’.
To briefly explain this contribution, I’ll start with the noun, and then move back to the adjective. ‘Experimentalism’ is a term that I have borrowed from European Union (EU) Studies which has begun to gain traction across Development Studies too.2 The concept of experimental governance refers to situations in which institutionalised processes of trial-and-error-based learning are used to lock operational improvement into a given governance system. I have been struck by the extent to which, from the commencement of its lending for housing, the Bank employed language and practices that are closely associated with experimentalist governance. Lesson learning, often driven by the emergence of evidence of operational failure, can be seen to have provided the building blocks of operational change in the Bank’s lending for housing, with periodic systematic reviews serving to complete processes of paradigmatic transition. Through these processes, flexibility has been maintained in the means used to secure housing system improvements, and the ends employed to define and evaluate operational success. Jacqueline Best (2016: 43) has recently suggested that institutional responses to operational failure tend to take the form either of conservative efforts to hedge against future failure or of more creative experimental adaptation. In its responses to challenges in the realm of housing policy, Bank practice has followed the latter, experimentally focused pathway.
This experimentalism at the World Bank has, however, not played out across a neutral terrain. A series of internal and external constraints have served to structure the space through which the organisation’s experimentalism has unfolded. Inside the Bank, changing personnel served to limit the realms of the possible within this experimentalist pathway. In particular, the removal through the mid-1990s of the small team of urban planning specialists from the Bank left lending for housing largely in the hands of financial sector specialists. With this shift, the focus on upgrading informal settlements and providing serviced plots for self-build activities largely dissipated from staff’s professional expertise.
Outside the Bank, changing preferences and political economy structures amongst its borrowers have shaped the organisation’s operationalisation of its lending for housing. It is widely acknowledged that international organisations are more effective at transplanting ideas and practices into contexts where ‘sympathetic interlocutors’ exist.3 In line with this dynamic, throughout the history of World Bank lending for housing, loans have tended to be distributed to governments whose commitment to supportive reform has been previously demonstrated. In the early phase of its lending for housing, World Bank loans aimed to strengthen government agencies’ capacity to provide improved housing; this agenda was operationalised with borrowers whose development plans through the 1970s and 1980s saw the state as playing a central role in driving forward processes of economic expansion. As Bank staff’s turn toward market-based provision was consolidated from the early 1990s, housing loans were signed off with governments whose reform agendas had begun to liberalise domestic political economy structures.
In concert, these internal and external factors have reinforced the direction of travel in the Bank’s trial-and-error-based learning towards the promotion of mortgage markets and private housing provision as the foundation stones of national housing systems. By mapping the contours of this constrained experimentalism, and paying attention to the interplay between these internal and external dynamics, through this book I develop a holistic vision of the drivers of change in World Bank lending for housing.
Beyond this contribution to literature on the politics of operational change in international organisations, two additional overarching insights are generated through this book. The first of these engagements is the challenge that is presented to dominant narratives on the post-Washington Consensus. Specifically, I suggest that the characterisation of the post-Washington Consensus as containing turns toward pro-poor targeting and institution building is not borne out by practice in World Bank lending for housing. The second of these engagements relates to the nature of contemporary state transformation that the World Bank’s lending for housing has helped to bring into being. In particular, I suggest that Bank loans have moved through the decades from supporting the consolidation of ‘providing states’ to supporting the consolidation of ‘regulatory states’. Existing literature on the regulatory state of the South raises questions over the potential for the achievement of progressive outcomes through increasingly marketised structures. Through this study, I suggest that the marketisation of housing systems has served to prioritise the needs of moderate- and higher-income groups, over the more pressing needs of lower-income groups.
These points of engagement with literatures on the politics of change in international organisations, the (post-)Washington Consensus, and the regulatory state of the South are outlined in the first chapter of the book, and further explored through later empirically focused chapters. In the following sections of this Introduction, I briefly flag the overall findings presented on the content of World Bank lending for housing, and the approach used to probe the interactions between the organisation and its borrowers.
I.2 HOUSING THE WORLD
Across the globe, there are thought to be around 1 700 000 000 individual human habitations.4 In Western Europe and North America, increases in mid-life spousal separation, low birth rates, and a tendency to avoid extended family living have seen the average rate of occupancy drop to just two people per household. In contrast, the 1 64 000 000 households located on the African continent each contain an average of five people. Across the global South, the number of people living in sub-standard shelter continues to rise. According to UN-Habitat data, the number of individuals living in ‘slums’ has risen consistently through recent decades: in 1990, some 650 000 000 people lived in such conditions. By 2000 the figure had increased to 760 000 000, and by 2012 stood at some 863 000 000.5
The material impact of housing standards on quality of life is difficult to dispute. For most individuals, whether renting or owner-occupying, the cost of shelter will represent the single largest expenditure made throughout their working life. The quality of shelter in which individuals reside will heavily impact on the patterns of sickness and health that they enjoy or endure, and can contribute to early mortality for both children and adults.6 The location of shelter will leave a heavy imprint on one’s education and working life; proximity to sites of learning, means of transportation, and centres of commerce and industry can limit the number of hours spent away from the home, and substantially raise earning power. However, in spite of its great significance, the issue of housing provision remains a relatively marginal feature within the field of Politics and International Studies.7
The World Bank has been lending for housing since the early 1970s, and since this time has committed around US$14 billion through some 240 interventions. However, this strand of operations has to date attracted a low level of attention. Existing works include Cedric Pugh’s (1994) evaluation of World Bank lending for housing through the 1980s and Cecilia Zanetta’s (2004) comparative study of operations in Mexico and Argentina, which tell similar tales of a disjuncture between the Bank’s grand modernist plans and more modest local capacity and need. More recently, Robin Biddulph’s (2014) forensic critique of the Bank’s land administration reform in Cambodia has shone a light on a case of elite capture, with in this instance lending for housing serving to aggravate existing socio-economic divisions. Through this book, I extend this small literature on the World Bank’s lending for housing by providing new detail on the content of changing practices in this area.
By coding World Bank housing projects from 1972 to 2012, I have identified three distinct paradigms of operational practice. The dominant form of operations during the first of these phases, lasting through the decade from 1972, consisted of supporting state agencies to directly provide improved housing. These interventions consisted of upgrades to existing low-standard accommodation, and the construction of plots of land with core utility connections for individuals’ self-building of housing. From the early 1980s, operational practice shifted decisively toward improving states’ capacity to provide finance for housing. From the early 1990s, this focus on national housing banks was itself superseded by the turn to private mortgage market expansion. Through this contemporary period, World Bank resources have been used to create subsidy schemes to expand access to commercial sources of housing finance, to directly pumpprime the sector, and to create mortgage liquidity facilities to link housing finance into secondary financial markets.
In addition to cataloguing these shifting modes at the heart of the Bank’s housing operations, I have also tracked other significant points of continuity and change in project design. By collating supplementary information on the nature of institution-building activity, forms of support for private sector construction and finance, and modes of pro-poor targeting, a comprehensive audit is presented of the changing shape of World Bank lending for housing.
Beyond the Bank-focused literature, the issue of housing finance has over recent years attracted some scholarly attention. Works exploring the intersection of housing finance and development have sought to identify the factors that inhibit mortgage market development, and to probe the distributional consequences that follow from the expansion of housing finance systems. Focusing on Latin America, Bouillon (2012: 189–210) pinpoints improved protections of creditor rights and enhancements to property registries as necessary conditions for housing finance expansion, and suggests that fixing these weaknesses could lead to improvements in the quantity of decent housing being provided. Focusing on East Asia, Doling and Ronald (2014) outline ways in which labour market regulation and government support for construction and credit growth helped expand rates of home building and ownership across the region’s major economies. While these institutional complementarities enabled the housing sector to become a key pillar of the economic growth models, Dorling and Ronald identify the division between owners and non-owners as a significant source of growing inequality across the region. From a more critical perspective, Soederberg (2014) has explored housing debt’s function as a form of market discipline over labour. Focusing on Mexico and the US, Soederberg suggests that state support for credit market expansion has enabled the expansion of this ‘debtfare’ compulsion.
Through this work, I demonstrate that the World Bank has for several decades been engaged in trying to lay the foundations for effective mortgage market expansion. Operations have focused on the relatively narrow types of institution building recommended by Bouillon, rather than on supporting the broader synergies between employment system, construction, and finance identified by Doling and Ronald. Whereas Soederberg explores the negative consequences of indebtedness associated with gaining access to housing credit, I highlight the tendency for the impact of World Bank-supported projects to flow increasingly to those on moderate and higher incomes – individuals who are able to access the formal circuits of housing finance that Bank projects have sought to establish and expand. In contrast, interventions have increasingly bypassed the low-quality shelter inhabited by poorer sections of society in borrowing countries.
I.3 THE WORLD BANK AND ITS BORROWERS
Much literature on the World Bank directs its attention to gauging the on-the-ground impact of the projects and policies supported by the institution (e.g. Rosenberg and Korsmo 2001; Szablowski 2002; Bradshaw and Linneker 2003; Bonal 2004; Dijkstra 2011). As noted above, existing works on World Bank housing policy have adopted this country case-study approach. A parallel literature has, however, sought to lift the lid on the institutional dynamics that lie behind particular processes of operational change in key arenas of global economic governance (e.g. Weaver 2008; Park 2010; Clegg 2014a; Gallagher 2014; Ban 2015). In bringing together these two focuses to explore the relationship between the World Bank and its borrowers, I follow the lead of works including Woods (2006) and Broome (2010b). By so doing, I aim to explore the interrelationship between Bank-level and domestic dynamics in shaping the process of operational change in this area, and also to present new detail about the impact of Bank projects on national housing systems.
To structure this exploration of the relationship between the Bank and its borrowers, I have drawn upon existing housing studies literature to develop a three-fold typology of national housing systems. Existing works have identified distinct Latin American, post-socialist, and sub-Saharan African modes of housing system.8 Here, I explore the changing nature of World Bank lending into each mode. I adopt the terms ‘mixed provision model’, ‘state socialist model’, and ‘informal model’ to refer to the three systems. The mixed provision model captures national housing systems that combine substantial state-led, market-based, and informal supply, the state socialist model captures systems characterised by public ownership and central planning, and the informal model captures systems with weak state structures and extremely low levels of market-based provision.
Across the three modes, overarching similarities can be discerned in the relationship between the World Bank and its borrowers. In particular, we see that the nature of housing system reform being promoted by the Bank has typically been complementary to borrowing-country governments’ existing policy agendas. This dynamic has held true within both the state-strengthening and the more market-oriented periods in World Bank lending for housing. Across Latin America and sub-Saharan Africa through the 1970s and into the 1980s, Bank loans helped support the expansion of national housing agencies’ operations at a time of broad-based government commitment to extensive state intervention. As debt crises and macroeconomic instability from the early 1980s served to catalyse the emergence of domestic reform agendas, World Bank housing loans provided a tool to enhance the role of private construction and private finance in this important sector. Across the state socialist mode, borrowing from the Bank commenced in the early 1990s. Here, the Bank’s emerging market-based paradigm matched with domestic governments’ search for tools to introduce and sustain non-state provision. While internally driven experimentalism has shaped the content of the Bank’s evolving approach to housing, domestic dynamics have provided receptive environments for the organisation to transmit its preferred operational practices into.
To explore the nature of the relationship between the Bank and its borrowers, I have followed a two-level approach. In addition to reviewing the broad sweep of experiences across each of the three modes identified, I have also explored one case from each mode in more detail. The individual cases studied have been selected on the basis of intensity of World Bank involvement. Mexico is focused on from amongst the examples of coordinated market housing systems, having received the highest number of Bank housing loans from within this cohort. China is focused on from amongst the examples of state socialist systems, and Tanzania from amongst the examples of informal systems. This case selection allows for holistic overviews to be presented of the interaction between Bank- and domestic-level dynamics, and for the impact of Bank projects on national housing systems to be tracked over extended timeframes. By combining the review of individual cases with the review of transformations across the modes from which these cases were selected, a comprehensive picture is presented of the relationship between the World Bank and its borrowers. In the paragraphs below, I close this Introduction with a brief précis of each of the three case studies.
In the Mexican case, the Bank’s state-strengthening paradigm was coincident with the Institutional Revolutionary Party (PRI) government’s agenda through the 1970s. In addition to fitting with the state-led approach to development, early Bank projects through the 1970s matched the ruling party’s use of social welfare as a ‘temporary palliative’ to ameliorate the negative outcomes caused by rapid urbanisation and industrialisation. Indeed, the first Mexican housing loan was specifically designed to rectify the urban chaos that resulted from the construction of a large steelworks in the Michoacan region.9 The first cycle of internal lesson-learning at the Bank saw, from the early 1980s, staff increasingly prioritising the reform of state-controlled housing systems. This Bank-side shift aligned with the rise within the PRI of the ‘técnicos’ cadre, who rapidly contracted large loans to ramp up the capacity of the National Trust Fund for Popular Housing (FONHAPO) to pass housing credit on to relatively low-income groups. From the early 1990s, Bank staff moved toward focusing on mortgage market expansion as the key tool of housing policy. However, it was only following the election of the National Action Party’s Vicente Fox at the turn of the millennium that the operationalisation of this market-based approach was consolidated in Mexico. Under successive Party of National Action (PAN) governments, the bulk of arrangements contracted with the World Bank have gone towards establishing and maintaining the Sociedad Hipotecaria Federal (SHF), a mortgage liquidity facility modelled on the US’s Freddie Mac.
With World Bank support, then, private finance and private construction have become an increasingly central feature of the Mexican housing system. However, even with the existence of mortgage subsidy schemes, the lowest-income 30 per cent of the population remain frozen out of formal circuits of housing finance. This 30 per cent includes the majority of the 12 million inhabitants of the country who continue to live in ‘slum’ conditions,10 with the most pressing need for improved housing. The Bank has found a like-minded policy elite with whom to operationalise its housing finance-based operational paradigm, and has helped support the move from a providing to a more regulatory vision of the state. However, over time the needs of lower-income groups have become increasingly marginalised in World Bank lending for housing to Mexico.
In the Chinese case, the start of lending for housing operations occurred with the confluence of the World Bank’s turn towards market-based mechanisms, and the consolidation of the country’s post-Mao liberalisation. Prior to the coming together of these dynamics, China’s housing system in the mid-1980s remained almost entirely dominated by municipal authorities and state-owned enterprises, featuring public ownership of housing alongside heavily subsidised rent rates. The initial loan agreements between the Bank and the Chinese government were aimed at realising the goals set out in the government’s 1988 Reform Strategy. The government’s Reform Strategy presented a commitment to transform housing ownership patterns; World Bank loans in 1991 and 1994 saw resources being released respectively to municipal authorities to construct housing units for private sale, and to non-state housing agencies to purchase existing public stock for commercial letting.
In the mid-1990s, the pace of reform was substantially ramped up. At this time the Chinese Communist Party’s State Council announced an extraordinarily ambitious plan to privatise all of the country’s state-owned housing stock, and to direct new-build constructions toward sale through market-based mechanisms. World Bank support has focused on efforts to expand the Chinese mortgage market, with technical assistance and training rather than large-scale financial investment becoming the favoured mechanism of intervention. Dramatic transformations have been achieved: from a negligible level in 2000, by 2012 around US$1.2 trillion of outstanding mortgages sat within the Chinese housing finance system. Throughout this period, the scale of private construction had also risen precipitously. However, as was the case in Mexico, the Bank’s lending for housing in China by and large has lacked effective mechanisms to target assistance to lower-income groups. Congruent reform agendas at the Bank and within domestic political spheres have been the achievement of wholesale transformation in the Chinese housing system; however, evidence of pro-poor impact from the Bank remains limited.
Through the third of the case-study evaluations, I explore the Bank’s engagements with Tanzania. After gaining independence in 1964, the Chama cha Mapinduzi government led by President Julius Nyerere strengthened state control over all areas of economic life. The Tanzanian National Housing Corporation (NHC) became established as a key pillar in the national housing system; the World Bank’s first housing operations in the country supported this state-led approach, aiming to bolster NHC capacity to directly provide improved housing. As the Bank’s approach became more market oriented and the Tanzanian state continued to extend its control over national economic activity, a significant hiatus in World Bank involvement in the Tanzanian housing system then followed. It was only following the election of Benjamin Mkapa to the presidency, and with him the emergence of a more market-oriented reform strategy, that lending for housing re-emerged.
Informal provision of housing has represented the dominant mode throughout the history of the Tanzanian housing system. World Bank resources in the 2000s sought to support the Mkapa administration’s efforts to expand formal private sector provision. While some focus has been placed on the supply of serviced plots and improvement to utility provision within informally settled neighbourhoods, the bulk of World Bank resources through the 2000s were channelled toward the establishment of the Tanzania Mortgage Refinance Company (TMRC).
It was in 2010 that the Tanzanian Mortgage Refinance Company was established, with the assistance of a US$30 million loan from the World Bank. Sixteen financial institutions signed up to become eligible to receive TMRC resources; by 2012, six of these members had on-lent World Bankderived resources, and steps had been taken by the Company toward the repurchase of mortgages for securitisation and resale into secondary financial markets. However, with housing credit remaining out of reach for all but relatively high-income individuals, the housing conditions of lower-income groups have not been impacted on by these flows of capital. While Bank-supported reforms in the realm of land reform and upgrading activities have had some pro-poor features, the direct benefit of the majority of resources released by the World Bank have not been effectively targeted.
In each of these three cases, World Bank lending for housing has gone with the grain of domestic reform programmes and policy commitments. While the content of operational paradigms has been shaped by trial-and-error-based experimentalism undertaken within internal World Bank networks, the operationalisation of these paradigms has remained contingent on the alignment of the Bank’s agenda with domestic institutions and preferences. These observations are broadly supportive of findings from existing literatures regarding the tendency of Bank staff to seek out ‘sympathetic interlocutors’:11 domestic policy elites whose frames of reference regarding the foundations of economic policy cohere with their own. The added insight generated here is that, while complementary efforts of domestic actors and the Bank may have helped catalyse transformations in national housing systems, these transformations have over time become decreasingly focused on the needs of the poor.
Following on from this Introduction, in Chapter 1 of the book I outline in detail the constrained experimentalist framework. In addition to clarifying the engagement put forward with literature on the politics of operational change in international organisations, I also outline engagements with literatures on the rise of the regulatory state of the South and on the nature of the (post-)Washington Consensus that are advanced though the work as a whole. In Chapter 2 I outline the Bank-side component of this story, providing a detailed account of the learning cycles that have shaped changes in the organisation’s lending for housing. I then in Chapters 3, 4, and 5 move on to the case-study-focused chapters. Here, I outline the coalescence of Bank- and domestic-level factors in the Mexican, Chinese, and Tanzanian cases, and the mixed provision, state socialist, and informal modes that these cases exemplify. Finally, the book closes with a chapter reviewing unfolding dynamics in World Bank lending for housing. While the strictures of constrained experimentalism mean that wholesale change is unlikely to occur, developments in the realms of housing micro-finance and water supply and sanitation present reasons for guarded optimism regarding the pro-poor targeting of World Bank lending for housing.