By Alison Wallace, Roger Burrows, David Beer, Alexandra Ciocanel and James Cussens. 

Tenant referencing is a critical aspect of accessing a home in the private rented sector (PRS). Referencing can be the decisive factor in decisions about who gets access to what housing. These decisions are now frequently arrived at  through a combination of the expanding availability of data and the processing power of algorithms. Recently, UK market regulators have raised concerns about the volume of data collected and the subsequent use of guarantors. The concerns are moving into the public realm, with, for instance, a  recent FT article expressing alarm at the depth of financial scrutiny of renters and the prospect of landlords requiring access to people’s bank account data. There is also growing concern about algorithmic decision making in accessing critical services, especially given the potential problems of exclusion and inequality in rental markets. 

Our Code Encounters research, funded by the Nuffield Foundation, looked closely at the use of the digital risk profiling tools that mediate access to housing across the market, including in private rental. The project illuminates contemporary tenant selection and the role of data and automation within it. In particular, we have identified the challenges of fitting complex lives into fixed algorithmic models, the importance of maintaining human intervention and the presence of intrusive data collection, compounding exclusion and raising issues with conditional lets. 

These digital tools for risk profiling sit at the intersection of proptech and fintech, drawing credit bureau and other data into affordability assessments and letting decisions. Existing evidence notes new data and automation supporting asset accumulation and tenant surveillance in private rental markets. In England, changing PRS regulation, welfare reform and the cost of living crisis have meant that landlords are increasingly actively managing their business risks, including through tenant selection. Choosing the right tenant prevents later issues and limits the prospect of having to bring tenancies to an end.  

‘So, it’s a bit hard to say but obviously, the removal of fixed-term tenancies is going to cause a problem, because people that are on a temporary contract, for example, do you let them in or not? If they’ve got a contract that ends in six months’ time, whereas before, you’d grant them a six-month tenancy, if I was a landlord, I would be declining them in the new world.’ (Tenant referencing firm 4)

Our interviews suggest that landlord and agents’ use of digital tenant referencing tools has grown, reflecting a shift from earlier analogue approaches to more algorithmically driven methods for evaluating tenants, with implications for housing accessibility and exclusion.

The increased availability of data reserves are now frequently applied to tenant referencing, aided by increased automation. These data sources are fast and convenient and include e-payslips, Companies House or HMRC data, fraud or anti-money laundering as well as traditional credit information, and these help verify tenants identity, income and other financial circumstances. In the case of Open Banking, access to detailed banking transaction data is framed as helping people overcome thin credit files and promote inclusion, although there were privacy concerns about the intimate nature of these data. 

‘Pretty much all the data, yes. That’s why I always thought open banking to be an extremely scary thing. ‘Give me your 11 months of data in half-an-hour, and I will tell you exactly what sort of a person you are in half a day.’ (TR2)

While data resources are expanding, there are significant digital data gaps relating to employment security and former landlord references in the UK private rental market that new data sources cannot overcome, and therefore limit the full automation of tenant risk profiling observed in other countries, notably the US and Australia. Many referencing services therefore adopt hybrid systems mixing human and algorithmic analytics and decision-making, although digital analytics of tenant circumstances are growing more sophisticated. 

Interviews indicate that many landlords and agents still place great value in qualitative data or tenants’ ‘soft attributes’ to augment formal tenant referencing and letting decisions, screening-out many tenants prior to formal referencing. Significantly, human intervention is also used to frequently override referencing recommendations, as landlords accept tenant explanations of weaker profiles or simply like the tenants. It is likely that more fully algorithmic tenant referencing would have greater exclusionary impacts without such human oversight. 

‘We could go to a fully tech solution, but actually, what we’ll do is just switch off a load of business.’ (Insurer 1)

The PRS is home to a diverse range of tenants but the tools struggle with complex tenant situations, leading to exclusion for some or significant human involvement in interpreting data and handling exceptions. 

Referencing platforms find people with limited credit histories, young people or migrants, those with adverse credit, and non-standard employment challenging, raising prospects of reinforcing existing inequalities. Squeezing people through fixed models that neglect geographical differences in market pressures and different landlords’ appetite for risk, the tools prompt greater use of conditional lets to cover risk, such as guarantors and rent in advance, that not all tenants are able to meet. 

‘I think that again, all the information that you are – how much you earn, this, that and everything, I don’t really think it should be relevant. I think it should only be, I think, showing that you can pay your past rent, and you have a rent book or something that states that you’re always, and your past landlord can say that you’ve always paid your rent. That should be sufficient, and it’s not any more. […] I think that’s changed loads over the years.’ (Tenant 15)

The drive for greater data use and automation among landlords and letting agents is clearly evident, incorporating digital risk profiling into end-to-end property management platforms, for example. However, there are no reciprocal data sharing arrangements in the PRS to test model accuracy, as in the financial services market, and little sensitivity to equality impacts. Tenants largely found the systems convenient to use but not all tenant circumstances could be accommodated in the  user interfaces. The greater use of Open Banking data may have benefits for those with thin credit files- younger people or migrants -but tenants often felt compelled to consent despite finding the practice intrusive. 

End-to-end property management platforms are growing, but digital rental futures may also incorporate two things. Firstly, public awareness and an evidence base lags Open Banking technology that is still developing at pace. Open Banking challenges people to present and manage their banking transaction data in the same way they do their credit scores, producing winners and losers. Secondly, the emergence of tenant passports, where responsibility for referencing is placed on tenants not landlords or agents, means a growing importance for how tenants manage and present their digital profiles, or as Fouhcade and Healey term their “eigencapital”. These systems favour the ideal professional tenant rather than marginal households who would be further disadvantaged in the marketplace, and become as foucade and Healey also dub the “lumpenscoretariat”. 

The research advocates for greater awareness, transparency and regulation of the volume of data collected and an emphasis on ensuring fair access to mainstream rental housing. 

The study was based on 122 in-depth qualitative interviews, 50 of which related specifically to private renting. The full report can be obtained from the Code Encounters website along with other outputs of the project.  https://www.york.ac.uk/chp/housing-markets/code-encounters/ 

For further information see Wallace, A., Beer, D., Burrows, R., Ciocănel, A. and Cussens, J. (2024) Digital tenant risk profiling in England’s private rented sector- Code Encounters Report 2. York/Bristol, University of York/University of Bristol.

To hear more about the wider Code Encounters project findings across the housing system you are welcome to join the online launch of our findings on Thursday 28 November 2024 9-10.30am. Please register for free attendance on this Eventbrite link.

You may also like…