Part 3

Chapter 12

The Credit Score Society

This chapter argues that the credit score — the most widely used algorithmic poverty measure — functions as a form of automated redlining that translates structural disadvantage into individual numeric ratings, regulating access to housing, employment, and financial services with no mechanism for appeal.

Drafting

Synopsis

The credit score is a welfare system that nobody voted for. It classifies people into eligibility bands, determines access to housing, employment, and essential services, penalises the poor for being poor, and is governed not by public law but by proprietary mathematics owned by three private companies. Bill Fair and Earl Isaac founded Fair Isaac Corporation in 1956 with an explicitly progressive argument: replace bank managers’ racialised personal judgements with a formula, and you will remove racial and class bias from lending. Within a decade the formula had encoded both racial and class bias. The FICO score’s feature weights — payment history 35%, amounts owed 30%, length of history 15%, credit mix 10%, new credit 10% — are not neutral measurements of creditworthiness. They are a political choice about which aspects of financial behaviour predict default, made by people who could not avoid bringing assumptions about whose behaviour was normal and whose was risky. The Fair Credit Reporting Act (1970), designed to protect consumers, inadvertently legitimised and standardised the credit bureau oligopoly in the act of trying to regulate it. The subprime mortgage crisis added a new mathematical layer: CDO tranching presented correlated risk as diversified, allowing the scoring infrastructure to generate profits from the pricing of poor people’s risk rather than their exclusion. The borrowers were destroyed; the scoring infrastructure survived. The credit score is now an eligibility instrument for housing, employment, insurance, and utilities — a private welfare state whose accountability is protected by trade secrecy and whose harms are, for the people it governs, practically impossible to contest.

In This Chapter

  • How Fair and Isaac’s formula for removing personal bias from lending encoded the same biases through its training data, and how the FCRA’s attempt to regulate the resulting system instead confirmed its authority
  • How CDO tranching used threshold mathematics to present correlated risk as diversified risk, converting the scoring infrastructure from an exclusion device into a profit-extraction device for the same populations it excluded from prime credit
  • How the score’s scope expanded from credit to employment, housing, insurance, and utilities without democratic authorisation, creating a private eligibility infrastructure that exceeds the welfare state in practical power over life outcomes for many poor people
  • How Sesame Credit’s extension of FICO logic across consumption patterns, social connections, and civic compliance demonstrates what the Western credit-scoring infrastructure becomes when the implicit becomes explicit
  • How “alternative data” scoring — rental history, utility bills, social media behaviour — is sold as financial inclusion for the “credit invisible” while functioning as the extension of the scoring apparatus into domains of life previously outside its reach

Connection Forward

Chapter 13 turns from the private scoring infrastructure to the British tradition of welfare respectability — tracing the single mathematical project from 1834 onwards that encoded the deserving/undeserving distinction as administrable criteria, and in whose most recent form — Universal Credit’s claimant commitment — both American scoring logic and Victorian Poor Law logic converge.

Key Claims