(music playing)

[MARION FOURCADE]

Welcome, everyone. I am Marion Fourcade, the director of Social Science Matrix, and I’m delighting to welcome you for this seminar of the computat– uh, of the Group on Computational Research for Equity in Legal Systems, or CRELS, which is headed by my colleague, uh, David Harding. Uh, so it’s a very special pleasure also to welcome back Julien Migozzi to Berkeley.

Julien is an economic geographer and assistant professor in the Department of Politics and International Studies at the University of Cambridge in the UK. Uh, after appointments at Oxford University and at the École Normale Supérieure. His appointment is specifically in development studies.

Uh, he works at the intersection of geography, urban studies, and economic sociology, uh, focusing on how, uh, digital s- technologies affect markets, cities, and inequalities, and he has a very specific interest in housing and financial markets that he’ll talk about today on he, you know, talking about his research in South Africa. I was actually privileged to host Julien, uh, as a postdoc. Uh, well, he was actually on a very, uh, leisurely postdoc from Oxford.

So Oxford was paying for the postdoc, but we got the benefit of Julien last year. He was here. Uh, and so many of you got to, uh, meet him and, um, and interact with him, so it’s really great to, to have him back.

And actually, during this time, he developed projects with Desiree Fields and with me. And, you know, so, uh, it’s been, uh, also, um, very productive for us. Uh, at Cambridge, he teaches a course on digital, uh, capitalism, and the big book right now that he’s, uh, uh, has been co-authoring, and I, uh, highly recommend that you check it out, is, uh, The Atlas of Finance, which came last year from Yale University Press and has been a hit around the world and is, you know, coming out in many, uh, many translations.

So without further ado, Julien, it’s your turn.

[JULIEN MIGOZZI]

Thank you.

(applause)

Um, yeah. Well, that thank you very much for this very kind, uh, introduction. Honestly, it’s, uh, it’s an honor, and it’s a pleasure to be back here.

I know it sounds kind of crazy to say that, but um, based on what happened in the US in twenty twenty-five, but I think I really truly had one of the best years ever being here. So I’m glad to be back here. So this talk is based on the, the work I’ve been doing in South Africa for the past ten years, and more specifically in the city of Cape Town.

I borrow from a recent paper as well as a book proposal. And, uh, with that being said, I’ll just jump, uh, straight in. Khayelitsha is Cape Town’s largest Black township.

The construction was started in the nineteen eighties to segregate thirty kilometers away from the city centre, populations classified as Black by the apartheid regime. At the time, it was illegal for Black people to own properties in urban areas, and mixed race people designated as coloured could only access homeownership in designated townships. Today, Khayelitsha counts between five hundred thousand and one million residents.

It keeps growing. Racially speaking, Khayelitsha remains hyper-segregated, but socially speaking, the township exemplifies how South Africa has experienced a shift from race to class as the basis of segregation. Some sections boast large houses with ambitious front gates, which contrasts with the large informal settlements that keep sprawling out around the township.

These shacks, where residents lack basic services, just as water, sewage or sanitation, embody the failures of the apartheids– the post-apartheid state to improve the lives of the urban poor. As you enter Khayelitsha from the main highway, you will be seeing a lot of ads, typically written or painted on walls. Car repair shop, driving schools, health clinics, cyber cafes, interface banking apps.

Most of these murals tend to be replaced over time as the economy goes. Last year, for the first time, I saw ads for crypto. One mural, though, remains untouched.

It was painted in twenty ten by the South African artist Faith47, and it reads, “The people shall share in the country’s wealth.” This is a direct quote from the nineteen fifty-five Freedom Charter adopted by the African National Congress, which promised that South Africa’s national wealth, land, mineral resources, and banks be restored to the people. In the distance behind the mural, you can see the shape of Table Mountain, and the city center of Cape Town elected the world’s best city to live in by Time Out magazine in twenty twenty-five.

Cape Town is known as the Mother City, the epicenter of colonialism. It is here that Dutch settlers arrived and seized the land. It is here that thousands of people saw their houses destroyed by the apartheid government when neighbourhoods were reclassified as white-only areas.

It is also here that you find today the most expensive neighbourhoods of the sub-Saharan continent, a coastal stretch of flats and villas that in recent years have been colonized by Airbnb as the city of Cape Town was rebranded as Africa’s tech capital by the flux of digital nomads and the multiplication of fintech firms. From where that mural was painted, the wealth stored in property and the wealth created by the digital economy feel very distant, like a promise that was never fulfilled. Thirty years after the end of apartheid, South Africa remains the most unequal country in the world, and South African cities, by any type of metrics, top the charts of income inequalities and racial segregation.

The racialized wealth gap engineered by settler colonialism and apartheid remains cemented in housing inequalities and property ownership. Eight percent of the popu-population owns seventy-two percent of the land, and the top decile of South African households own sixty percent of the total housing wealth. But don’t be fooled by this feeling of inertia, because market forces are at work.

If you drive a little further into Khayelitsha, you reach the premises of Zolam Properties, a local agency founded by Zola, one of the very few Black property practitioners in the city of Cape Town. I have been coming to Zolam Properties since 2015. Most of their clients buy in the Khayelitsha area or seek to move out of the townships across the highway.

Like any real estate agency, ZOLAM Properties spend a lot of money on tech tools in the form of subscription. A platform to post their listings, a platform to download automated property values, a platform to manage their small rental portfolios. Over the years, this agency offered me a unique window from which to observe how technology shape a market in the making and how property dynamics intersect with class formation.

And this is how I met Michelle. Michelle is originally from Zimbabwe. She’s smart, she’s fun, and she’s a successful property broker.

She knows the home buying process inside the, inside out, and she can scan a client in a one minute. Michelle is particularly good with social media. With her TikTok account, Property Lioness, she has adopted the latest marketing trend where agents uses live video to advertise a property.

TikTok works well for Michelle, even though it takes a lot of time and work to create the content. Of course, not every property that Michelle is advertising gets this digital exposure. Michelle is careful to curate her TikTok account.

On principle, she only shows property located outside Khayelitsha, where the streets look nice and safe, and the property seems in good condition. Not long ago, she was showcasing a house in Blue Downs, a neighborhood opposite Khayelitsha. A lot of people reacted, but one of the comments caught my attention.

That wa– That comment said, “What should your credit score be, please, in order to apply?” As her answer shows, Michelle is very knowledgeable about credit scores and how these affect applications for a mortgage. She deals with credit scores every day.

She explains to clients how important they are. She knows that at Zolam Properties or agencies, mortgage– seventy percent of mortgage applications get rejected because of poor credit scores. So thirty years after the end of apartheid, having a good credit score to get a mortgage and call a place home has become a social and organizational norm.

Proprietary market-based algorithms are used every day to frame market interactions that regulate the access to and the distribution of property wealth previously supervised by the state through the enforcement enforcement of racial categories. The work of Michel and the use of credit scores led me to question how digitized property markets and automated classifications affect wealth inequalities and class formation in an emerging, deeply unequal economy. And to contextualize this question, I’d like to highlight here three key elements of consideration.

The first one is the global decoupling between income and property prices, a trend that South Africa experienced like many other countries. This evolution prompted a few scholars to revisit our understanding of class formation and foreground the roles of asset ownership for the making and the future of twenty-first-century inequalities. As Lisa Adkins and Melinda Cooper asked, what becomes of class when residential property prices in major cities around the world accrue more income in a year than the average wage worker?

With the stagnation of salaries, the casualization of labor, and the inflation of housing prices, property ownership challenges and even trumps the traditional understanding of stratification based on income– based on occupation and wages. My second point is that this interest for landed assets for property markets actually creates a conceptual and empirical connection with the work of researchers that have invested– investigated social change and middle-class formation in the Global South. Since real estate represents the bulk of the middle class’s assets globally, and to overcome the limits of definition based on income or occupation that Bourdieu called statistical artifacts, scholars turn to real estate markets.

Since housing is both a home and an economic asset, a source of material and cultural capital, housing markets as a sphere of government interventions, a site of political struggle, and a vector of spatial segregation offer a good ground to theorize and document the middle class’s boundary work, which designates the distinction formed through interest, lifestyles, values, aesthetics, and socioeconomic positions. Claire Mercer, in her recent book centered in Tanzania, shows how property investment is central to middle-class formation, demonstrating that suburbs and middle class are mutually constitutive, as households access the right kind of house in the right kind of neighborhood. To center asset inequalities, class formation needs therefore to be understood as a social-spatial process of distinction.

The third element I want to highlight is the key role of digital technologies in mediating contemporary housing markets. With the rise of the platform as a business model, the datification of everyday life, and the digital mediation of transactions, property markets have entered the era of automation and big data. Algorithms support the calculation of property values, the visibility of listings, the underwriting of mortgages, the determination of insurance prices, the screening of tenants, the allocation of investment capital, and the management of properties.

As Sara Safransky recalls, property requires constant doing. In the last decade, cities have been confronted to the rise of a newish, boyish, hybrid, data-driven industry called PropTech, which seeks to automate and profit from the relationship and practices of landed property. Yet recoding property markets from scratch is, of course, impossible.

As evidenced by the work of James C. Scott on the construction of the cadastre, we should understand any attempt to reconfigure property markets and property systems as a a project of social engineering. So how can we document and theorize class formation from the lenses of stratified, financialized, and digitized property markets? And can we do this not from the Silicon Valley Wall Street nexus, but from the context of South Africa?

Since class is relational, understanding middle-class formation requires to unpack the market structures that regulate the production of and the access to housing assets, and contextualize the boundary work of the middle class within these wider logics of social certification. So this is my goal for the rest of the talk, exploring market structures to locate the mechanisms of middle-class formation as a social spatial process. Summarizing the history of property and inequalities in South Africa is an impossible task.

I will have here to sacrifice, to some extent, the social complexity of settler colonialism, apartheid, and post-apartheid. But to keep it simple, what South Africa shows in terms of property system, market-making, and class formation is a transition from the law of race to the law of the market. For centuries, the recording of property rights was structured on racial classifications that encoded the economic interests of the white settlers.

The legal and technical definition of private property was instrumental to code land into capital along a racial divide. Since Dutch settlers imported the, imported the cadastre in the Cape Colony, the population has been classified into racial categories endorsed with unequal tenure rights. In nineteen thirteen, the Union of South Africa passed the infamous Natives Land Act, which left only seven percent of the land to native populations.

The making of race and the making of property were a joint and racist process of social and spatial engineering. The apartheid regime then expanded the suite of legal tools that allowed colonial elites to take control of urban real estate through evictions and planning. Neighborhoods were classified into racial categories and destroyed or built accordingly.

Interestingly, apartheid was conceived as a very modern project with a great belief for the ordering capacity of technology. Government and banking industries were among the first in the world to adopt computers, catering for the apartheid mania for measurements. IBM computers, in particular, allowed them to punch the number of ID cards, where one digit would indicate the race of the cardholder.

So through the state-driven classification of people, neighborhoods, and properties, apartheid imposed a racist control of the housing market. The racial classification of the buyer, as recorded on the title deed, had to match the classification of the neighborhood. Banks, therefore, were mostly in charge of fostering white homeownership while redlining townships.

Apartheid engineers and policymakers placed strong emphasis on the power of data and statistics to optimize the economic system built on cheap Black labor and systemic racial segregation. And to find evidence of that, you need to, you need to look no further than the UC Berkeley Library. This book was printed in 1960 by the South African Council for Scientific and Industrial Research.

As indicated by the title, A Survey of Rent-Paying Capacity of Urban Natives, the goal was to construct a data-driven knowledge of the lives and finances of black households who, legally excluded from homeownership, had to rent in the state housing stock. So through the modeling of itemized budget expenses, the population was made legible to the gaze of the state and categorized into various brackets. The ultimate objective was to identify the level of rent that would optimize rates of collection in order to design future townships.

In the nineties, the law of race as the primary mechanisms of urban segregation and market control was replaced by the law of the market. Legally, the right to housing has been enshrined in the Constitution. Racial restrictions of their tenure, location, or property acquisition were removed.

The right to buy or rent became only regulated by strict affordability. You live whenever you can afford. A symbol of exclusion and white privilege, homeownership became a social pillar of the post-apartheid housing policies.

The state embarked upon a large-scale construction program to facilitate the access to ownership for the poorest of the poor. But by turning homeownership into a pillar of the new social contract, the state also sanctified private property. And despite the conspiracy theories promoted by Musk or Trump, the redistribution of landed property in South Africa remains a legal illusion, especially in cities.

As the state withdrew from the planning of segregation, from the control of the markets, leaving untouched the property inequalities engineered by apartheid, it means that for people to share in the country’s wealth, it has to be done through the medium of the market. And for mortgage lenders and real estate agencies, the rules of the game changed. The historical market makers, the historical intermediaries of housing wealth, had now to cater for the country’s majority.

The government made it clear that they were expecting banks to make mortgages available to populations historically excluded from housing finance or included on predatory terms. A whole new market had to be built. But over the last twenty years, the law of the market aggravated actually the housing wealth gap.

Through a partnership with the City of Cape Town, I was able to access twenty-one million of titled deeds that go back to the arrival of Dutch settlers in the seventeenth century. I also reconstructed streets by street, neighborhood by neighborhood, the racial zoning of apartheid to locate every property transactions in the racialized social hierarchy of neighborhoods engineered by settler colonialism and apartheid planning. Using a mix of computational and spatial analysis, I then turned these twenty-one millions of title deeds into a geolocated longitudinal database of property transactions that covered the whole metropolitan area for the past thirty years, allowing me to track the evolution of housing prices at the, uh, house level.

And so during the economic boom of South Africa in the early 2000s, as you can see, house prices skyrocketed between twenty– 2003 and 2007, replicating this global trend of housing price inflation. But as you can also see on this graph, housing prices, this inflation of property values was heavily racialized in Cape Town. The distributional effects of capital gains primarily benefited white owners, increasing the equity gap between formerly white-only areas and the rest of the city, building effectively a wall of money around the upper class, formerly white-only neighbourhoods.

The other consequences is that the decoupling between housing prices and salaries reinforced mortgages as an obligatory passage point towards homeownership for the asset-deprived minority excluded from the inflation of property values. With the declining share of the white population, strong rates of urban growth and social aspirations for home ownerships, banks were legally compelled, but also financially motivated to extend the frontier of mortgage markets. There were also external factors.

South African banks wanted to enter the global securitization markets organized by Fannie Mae and Freddie Mac, and they needed to standardize their lending policies to, for that, for that purpose. And so this is where credit scoring came in super handy as a seemingly color-blind technology to standardize risk management and make life-impacting decisions on the people and areas once discriminated on the basis of race. In South Africa, it only takes an ID number to access in a few seconds a detailed 15 pages credit report on anyone living in the country.

The central role of credit scores in today’s property market is due to legal and technical reconfigurations. The mania for measurement that characterized apartheid easily survived the transition to democracy. In the late nineties, the South African government became the global pioneer for the adopt, for the adoption of biometric technologies.

The skeleton identification system of apartheid was upgraded for the digital era through a new smart ID program. Meanwhile, South African turns in mass to credit-driven consumption. Retail stores aggressively targeted the townships, sending personalized, ready-to-use credit cards.

As people used credit to survive and to finance social aspirations in what Gillian Hart described as a jobless form of growth, the debt of households exploded. In the words of Deborah James, “A startlingly racialized debt became the structural feature of the post-apartheid society.” And mass-driven credit consumption resulted in the unprecedented process of datafication of the South African population, adding a flesh of consumer data to the existing informational skeleton with a level of granularity that the apartheid state would have dreamed about.

More importantly, in 2007, the National Credit Act requires now any lender, including banks, to perform a credit check in order to prevent reckless lending, which further pushed the creation of a consumer data market and the generic use of credit scores. Today, each use of a credit card, each late payment leaves a digital trace that can be harvested across various sites of consumption and citizenship through the national credit reporting system. In a short time span, South Africa became a goldmine for data brokers and credit bureaus when consumer data circulates across a wide information dragnet linking thousands of data points through ID numbers.

With such abundance of data and historical appetite and technical capacities for classification, South Africa became a testing ground for increasingly sophisticated algorithms to classify mortgage applicants. Classifications that take into account the residential location of the applicant, therefore baking in racial segregation into the classification of people. Not only do credit scores qualify or disqualify mortgage applicants, they also determine the interest rates and the loan-to-value offered by the bank.

As this home loan consultant told me, the cr- the credit matrix is the Bible by which they work. Discourse, of course, bears the watermark of existing inequalities, with consumer data being imprinted by the legacy of urban segregation and labor exploitation.

In twenty seventeen, seventy percent of over-indebted consumers were Black African, and the current rejection rate for home loans stands at about forty-five percent. As such, a seemingly colorblind, legally enforced digital technology reproduces at scale, racialized market outcomes in terms of residential mobility and access to property wealth. But how do these organizational practices and their stratifying effects translate into the spatial structure of the property market?

How can we mac– Uh, sorry, how can we map the work of algorithms when credit scores are proprietary data and impossible to access? That’s where I use the property database that I created during this research.

And that’s where it gets a little bit technical, but I’ll keep it very simple. For each transaction in the database, I compute a credit to price ratio. I calculate this ratio by dividing the mortgage amount by the property’s selling price.

And this metric captures the relative weight of debt, the debt involved in the purchase of a property asset. A ratio of one indicates that the property was fully financed through debt, meaning no deposit, no asset was required. A ratio of zero corresponds to a full cash-based transactions with no reliance on credit to acquire homeownership.

And the values in between reflect varying degrees of leverage. Higher ratios indicate that the buyers rely more heavily on credit to complete the purchase, whereas lower ratios suggest that the buyers have to contribute with a larger deposit and depend less on borrowing. And then I used spatial interpolation method to map this variation of the credit to price ratio at the neighborhood level.

And this allows me to identify different types of housing submarkets created and shaped by the unequal distribution of prices, mortgage lending, and financial capacities. And the resulting map of the credit to price ratio needs to be analyzed and compared with the racial zoning of apartheid on the top left, and the geography of housing prices on the bottom left, which are, of course, heavily correlated. It shows unequal regimes of debt, property– debt-driven, sorry, property acquisition.

Where the map is blue or dark blue for the most expensive neighborhoods, prices are too high to get a mortgage, so buyers have to leverage existing housing equity, savings or family assets to purchase property. Yellow areas, on the other hand, show neighborhoods where the property market is completely dependent on mortgage policies and mortgage lending. Acquisition of property assets and residential mobility in these neighborhoods are conditioned to mainstream financial products regulated by credit scores.

And this area, precisely here, named as Blue Downs, is where Michelle works as an estate agent. And this is where, for the rest of the talk, I would like to locate and flesh out the boundary work that shapes middle-class formation. So located in the eastern suburbs of Cape Town, Blue Downs is a mortgage-shaped, mortgage-dependent market.

It is situated on a spatial corridor of upward mobility that received middle-income households willing to move out of the townships of Mitchells Plain and Khayelitsha. Since the late 2000s, private developers have driven a sporadic urban sprawl, producing new housing stock in the form of gated estates scattered across grazing farmland. And from that space, I suggest understanding the middle class as a filtered class comprising asset-deprived households that manifest their boundary work by navigating the market to access debt-leveraged homeownership.

And I identify three mechanisms to middle-class formation. First, they can afford to pay up mortgages and tran– and transaction costs. Second, they score high enough under the market’s automated gaze.

And third, they settle in gated enclaves. Crucially, algorithms are involved at every steps of the three mechanisms in both hidden and prominent ways, shaping the political economy of housing and data. In Blue Downs, the market is driven a hundred percent by mortgages, so banks decide who moves in, who buys or not.

Conservative lending practices structured upon high interest rates mean that mortgages are unaffordable for eighty percent of the population. Only the upper quintile of households earn enough to be considered by banks. To calculate the affordability of customers, banks source credit bureau data, consumer data, and deduct existing expenses to reconstitute a proxy of disposable income against which future mortgage repayments are calculated.

Local agents note that when a mortgage offer, when a mortgage offer requires a small deposit, many aspiring homeowners can be eliminated. They may qualify in terms of income, but they live paycheck to paycheck. So, uh, ninety percent or eighty-five percent loan-to- value eliminates them from the home buying process.

Additional costs like transfer duties, attorney fees, bond registration also adds to the affordability challenge that we need to unpack beyond the sorting mechanism created by housing prices, wages, and banks’ lending policies. To afford mortgage homeownership, households must mobilize, must display various forms of economic and social capital, exercising boundary work by investing time, money, and energy to navigate the market and make themselves legible to state and financial entities. In the context of mass unemployment and systemic poverty, even basic market interaction, such as finding a house, attempting to do a viewing, providing legal documents, reflect distinctive property practices.

Scrolling down on a South African Zillow costs a lot of money to have to get the data. Organizing a viewing requires dealing with poor public transport, heavy traffic, and limited car ownership. And estate agents are interviewed, described how just gathering documentation for a transaction is often complex and time-consuming.

Engaging with a bank demands legal literacy. You need to gather, scan, and submit ID documents, bank statements, payslips, marriage certificates, proof of employment. It might be routine for a few, but it’s a si-significant hurdle for others.

Executing a transaction also requires building and maintaining a relationship with key intermediaries such as estate agents, attorneys, and mortgage brokers, often over several months, during which unexpected costs like car accidents or medical bills can disrupt the household budget and derail the transaction. Understanding all the components of affordability emphasizes how being able to engage in a property transaction manifests boundary work in terms of lifestyle and socioeconomic status. However, even for those who can afford a mortgage, scoring high, scoring high enough remains essential.

Performing credit checks to sort customers. It’s a daily task for agents in Blue Downs, especially the ones working in gated estates. In the context of high indebtedness, high rejection rates, with sales target to meet and extensive paperwork, agents prioritize clients who have higher chances of mortgage approval.

And to that end, they want to sort customers as fast as possible. So they use software provided by market leaders in mortgage origination, such as UBA or BetterBond. They enter the ID number to perform a real-time credit check.

Advances in cloud computing and machine learning have accelerated and obscured the classification process. In 2017, credit checks were conducted on laptops. In 2024, I saw an agent using her phone to perform facial recognition to screen in walk-in customers.

The app would scan the face. It was developed by a new PropTech company, and it connects to the biometric databases of the Department of Home Affairs and then retrieves information from credit bureaus. Within three minutes, in front of my eyes, the following brief email was received.

“We have conducted a credit check for name. Their status is red. This client needs to work on his score.

The BetterBond team will be in touch with you. Since the exact score is not communicated, agents rely on color-coded categories: green, yellow, amber, to guide their decisions. Only green and amber profiles were processed by the sales agent.

So the use of credit score has significant market outcomes. Low score results in agents not serving certain clients. In these estates, clients classified as red, interpreted as clients that needs work, were offered instead rental options.

Low score also results in higher interest rates and mortgage payments, increasing the costs of housing. So under the market’s algorithmic gaze, acquiring property assets require being legible to the state, but also, as recalled by Marion Fourcade and Kieran Healy, available for measurement. And what I really want to emphasize here is that credit scores intervene at many stages of the transaction process.

Agents check clients first, but then banks do it again later to evaluate the mortgage application. And they do it again later when the bond is registered. So scoring high is not a one-type performance.

It needs to be repeated consistently over time during the whole transaction process. And this should be seen as a critical form of boundary work in digitized and financialized markets, even if it’s not necessarily conscientized. Because the calculation of credit scores draws on various aspects of economic and administrative life: employment, income, consumption, residential consumpt– residential location.

So credit scores weave algorithmic rankings and moral judgments with two years of data being sourced to calculate them. And in this context of acute social scoring, The boundary work of middle-class formation results in gating in. Results in the production of a stratified suburban space where the physical gates and the algorithmic barriers that surround these gated estates materialize class boundaries in terms of assets, aesthetics, and relationship to private property.

Crucially, for the relationship between property wealth, capital gains, and middle-class formations, properties located in these gated estates record a systematically higher housing appreciation compared to the rest of the area. This is due to strong demand and good marketing, certainly, but also because the appraisal algorithms used by the banks to determine property values always gives a positive weight to the presence of a security complex. Ubiquitous in Blue Downs, gates, walls, and fences serve both economic and marketing purposes.

Developers, aware that the clientele comes from townships associated with high crime rates, replicate a similar gated model to signal lifestyle and security. However, of course, gates vary in quality and effectiveness. One of the site managers admitted to frequent break-ins.

The true significance lies in the gate symbolism, representing secure estate living as a marker of social distinction. Gated estates also bind residents, as you can see on the top, uh, picture here, to distinctive property relations to safeguard the collective value of property assets and maintain a contrast with nearby townships. Homeowner associations enforce strict articu- architectural guidelines, leaving little room for individual differentiation in housing aesthetics.

They also try and regulate the social norms, such as the on-site behavior of residents in terms of noise or presence of guests. For the residents I interviewed, navigating the high cost and lengthy process of debt-driven acquisition through lifelong mortgage repayments yields not only the perspective of slow wealth accumulation, but also immediate social distinction behind the gates. And since class is fundamentally relational, I want to conclude now by mapping out, fleshing out the relational properties of the South African middle class produced by the sorting mechanisms of the market.

Shaped by centuries of land dispossession and deprivation of property rights, aspirations for a property-owning democracy in South Africa have been undermined by structural unemployment, jobless growth, asset inequalities, and prolonged economic stagnations. Racialized and selective lending practices persist today via mass datafication and algorithmic classification, shaping market structures and market outcomes. The relationship to property and debt is therefore a distancing feature of middle-class formation and boundary work.

The capacity to attempt to successfully navigate market structures distinguish the middle class from the urban poor and from the elite. Unlike the urban poor who access housing via squatting or through government programs, middle-class households are rather filtered in by the real estate and financial industries. Unlike the urban poor, they are calculated as eligible and incorporated into the bank’s mortgage portfolios.

They are not banished by market forces, by bulldozers, but they may be evicted if the sheriff, by the sheriffs, or if they default on their mortgages. This capacity also distinguishes them from the upper class and the elite. The acquisition of property assets is exercised under heavy economic constraints produced on the one, on the one hand by the legacy of racial property regimes and on the modern credit-driven, credit score-regulated construction of mainstream financial products.

Unlike social elites, middle-income households have limited purchasing options, and their locational strategies are restricted not only by social preferences but by the racialized distribution of housing wealth. So to access and accumulate housing wealth, the middle class, unlike the urban poor, doesn’t wait for the state, but for the banks. However, unlike the upper class, market interactions with financial institutions are largely depersonalized and automated.

The opportunities to explain a poor score to negotiate lending terms through private branking– private banking, sorry, are, of course, not available. Retrenched behind computational walls, banks communicate their decisions via minimalist PDF letters that state that the application has been declined on score or does not meet the minimum requirements of our credit scorecard. In the digitized market, the rules are clear.

Scoring high enough is a prerequisite. Upper-class buyers may receive a low credit scores, but they possess sufficient assets and savings to secure property ownership or offset a lower loan-to-values. Conversely, a low-income individual mets may score high enough through supervised consumption, but is more likely to be just excluded from homeownership by the sheer cost of housing.

So scoring high enough under the market’s algorithmic gaze attests to a distinctive socioeconomic position, consuming but avoiding over-indebtedness, budgeting, finding housing opportunities, scoring high enough many times, applying for a mortgage, waiting for the bank’s decision, moving in and getting in by committing to lifelong repayments, avoiding repossession in a stagnating economy. The housing market is a crucial arena where middle-income households earn their middle-class stripes. And as housing wealth and digital technologies increasingly shape social spatial inequalities, South Africa offers a stark illustration of how urban change, automated classification, and value creation become intertwined, bound together through the interplay of physical and algorithmic barriers.

Under digital capitalism, the enclavization of urban life and the rise of algorithmic social sorting go hand in hand with the assetization of data. In places like Blue Downs, this dynamic of social engineering and value creation becomes strik– strikingly tangible. For any gated estate, detailed granular market reports on residents, their personal finances, and their mortgage properties can be purchased online for as little as five dollars on credit.

Thank you very much for your attention.

(applause)

(music playing)