Synopsis
The postwar welfare state did not abolish the arithmetic of power — it reorganised it. In London, Beveridge’s flat-rate contribution formula replaced the Poor Law’s moral assessment with actuarial calculation, turning the claimant from a supplicant before a magistrate into a parameter in a formula. In Stockholm, Alva Myrdal’s Nation and Family proposed a different formula for a different person: earnings-related benefits designed around a dual-earner household where women’s paid employment was an actuarial assumption, not an exception. Both models shared the foundational mathematical move — parameterisation at population scale — and both were genuine progress over what came before. But they encoded different political visions of who the standard person in the formula was. Beveridge’s was a male industrial worker with a dependent wife; Myrdal’s was a household of two economic actors whose childcare and parental leave were public responsibilities. The contrast reveals that the gender politics embedded in each welfare arithmetic were not technical necessities but political choices — choices whose consequences persist in the administrative machinery that algorithmic governance would later inherit.
1. Two Tables, Two Arithmetics
In the winter of 1942, two documents were shaping the postwar welfare state. In London, Beveridge’s Social Insurance and Allied Services was published on 1 December to queues outside His Majesty’s Stationery Office — 635,000 copies sold of a government white paper, a figure that should still astonish. In Stockholm, Alva Myrdal’s Nation and Family was circulating among the policy intellectuals building the Folkhem. Both proposed to replace the Poor Law’s moral inquisition with actuarial calculation at population scale. Both treated the welfare state as a mathematical instrument: contribution records, benefit schedules, eligibility thresholds, and projected claim rates that could be calculated and administered without reference to the character of the individual claimant. And both embedded, in the apparently technical details of their formulas, a political vision of who the standard person at the centre of the system would be. They chose differently.
Beveridge’s standard person was a male industrial worker. His flat-rate model — a single weekly insurance stamp purchasing a standard package of benefits — expressed a democratic ideal: social citizenship as equal membership in a shared risk pool. The contribution rates and benefit levels were derived partly from Rowntree’s nutritional surveys and partly from Government Actuary projections. They carried the authority of calculation rather than of political negotiation. But the actuarial neutrality was incomplete: the 40-shilling couple rate was not 24 shillings for her and 16 for him. It was 40 shillings for a household unit defined by his employment. Nancy Fraser’s analysis of the gendered two-track welfare state names this structure precisely: a masculine subsystem of social insurance tied to labour-market participation, alongside a feminine subsystem of public assistance addressed to those outside formal employment. Beveridge’s universalism was built on the first track.
Myrdal started from a different question: how to engineer a society in which women’s paid employment and motherhood could coexist. Sweden’s total fertility rate had fallen from over four children per woman at the century’s turn to under 1.7 by the early 1930s, and the Myrdals had reframed this not as moral decline but as a rational response to structural conditions — inadequate housing, insecure income, the impossibility of combining paid work with childcare. The policy instruments that followed — universal child allowances, publicly funded childcare, school meals, maternal and child health clinics — were justified in the same actuarial terms that Beveridge used for National Insurance: projected costs, anticipated returns, population-level sustainability. The mathematics was the same kind. The person at the centre of it was not.
2. From Moral Failure to System Parameter
The Poor Law asked one question of the person standing before it: do you deserve relief? The answer was personal, contingent on the judgment of a magistrate who saw the applicant face to face. The 1834 New Poor Law formalised this into the workhouse test, but the logic remained individual and moral. T.H. Marshall noted that under this regime paupers “suffered a species of attainder” — they ceased to be citizens in any true sense while receiving relief. The price of support was the forfeiture of social standing.
Beveridge’s scheme replaced this question with a different one: has this person accumulated sufficient contributions to qualify for benefit? The shift was from moral assessment to actuarial record. Instead of a magistrate evaluating character, a clerk checked a contribution card. The contributor was entitled to benefit in the same way that an insured person is entitled to a payout: by the terms of the contract, not the judgment of the provider. Marshall’s culminating claim — that social citizenship was the twentieth century’s addition to the civil and political rights of the previous two centuries — rests on exactly this shift. Rights-based entitlement replaced discretionary relief.
But the actuarial model performed its own kind of abstraction. The claimant ceased to be a person with particular circumstances and became a parameter in a formula. Ian Hacking’s analysis of administrative classification is directly relevant: categories do not merely describe pre-existing kinds of person. They bring new kinds into being. The welfare state’s contribution categories — contributor, claimant, dependent, insured person — were administrative constructions that created institutional identities, with specific entitlements and specific relationships to the state, that the older vocabulary had not contained. A woman whose domestic labour produced no contribution record was categorised as a dependent — not because she did not work but because the system was not designed to record the work she did.
3. The Folkhem’s Mathematics
ATP — Sweden’s earnings-related pension system, introduced in 1960 after a political battle that ran through most of the previous decade — is the dual-earner welfare state’s defining actuarial instrument. Before ATP, white-collar workers had occupational pensions; blue-collar workers organised through the LO unions largely lacked comparable provision and faced a sharp drop in living standards at retirement. The Social Democrats and LO argued that a statutory earnings-related scheme was the only way to extend middle-class security to manual workers without creating a two-tier society in old age. The ATP bill passed the Riksdag’s Second Chamber in March 1959 by a single vote, when liberal MP Ture Königson abstained rather than oppose it.
ATP’s formula encodes a different actuarial logic from Beveridge’s flat-rate benefit at every level. A full pension required thirty years of covered labour-market participation; the pension base was the average of the fifteen highest earnings years; a full ATP pension replaced sixty per cent of that base. The system was formally gender-neutral: there was no dependent-wife category, no derived benefit, no contribution made on behalf of a household unit. But formal gender-neutrality in a labour market that was not gender-neutral produced gendered outcomes. Swedish women’s employment in the 1960s was concentrated in part-time work and lower-paid sectors; a woman who had worked part-time to manage childcare accumulated pension points for her actual hours and wages, not for the full-time equivalent the fifteen-best-years rule favoured. The formula had a standard person at its centre: a worker with a continuous, full-time, adequately paid thirty-year career. That worker was a different fiction from Beveridge’s male breadwinner, but it was still a fiction.
4. The Statistical Infrastructure
The postwar welfare state required, and assembled, a statistical infrastructure that no previous government had possessed. The Beveridge Report’s actuarial projections depended on estimated sickness rates, projected unemployment levels, and demographic trajectories derived from the General Register Office and the Government Actuary’s Department. The contribution record system created a longitudinal database of individual economic activity — a time-series of stamps, covering periods of employment and interruptions, available for retrieval when a claim was made. The Family Expenditure Survey, begun in 1957, and the General Household Survey, from 1971, extended the state’s informational reach from the individual contributor to the household as a unit of observation.
This infrastructure was assembled to serve the actuarial welfare state’s administrative needs. It was also the infrastructure that later algorithmic governance systems would inherit. Universal Credit’s real-time information system, which draws earnings data directly from HMRC’s PAYE records, is the Beveridge contribution card’s digital descendant — the same principle (track individual economic activity, derive entitlement from the record) executed at computational speed and granularity Beveridge could not have imagined. The political choices embedded in the 1940s infrastructure — which kinds of work count, whose labour generates a record, which household configurations fall inside or outside the model — persist in the systems built on its foundations.
5. The Scottish Settlement
Scotland did not have one relationship to Beveridge’s welfare model. In the industrial Lowlands — Glasgow, Motherwell, the mining villages of Fife, the shipyards of the Clyde — the male breadwinner in regular full-time industrial employment came closer to the system’s standard person than almost anywhere else in Britain. In the crofting counties — Sutherland, Ross and Cromarty, the Western Isles — the standard person did not exist. The crofting household combined subsistence cultivation with seasonal wage work, fishing, and employment in multiple income sources in ways the actuarial model could not track. Scotland before Beveridge had been running a distinct poor relief system since the Poor Law (Scotland) Act 1845: a system built on intimate local knowledge, resistant to standardised administration, that permitted outdoor relief and whose Board of Supervision’s records document a century of case-by-case negotiation between Edinburgh and parish boards that would be entirely foreign to the National Insurance clerk checking a contribution card.
The National Assistance Act 1948 abolished the Scottish Poor Law as part of the UK-wide welfare settlement. But the mismatch between Beveridge’s model and the crofting economy was structural, not accidental. The National Insurance unemployment benefit rules were built around the concept of a “full normal extent” of employment — benefit was not available for days when a worker “does not in the normal course expect to be working.” For crofters and Highland workers whose normal pattern was precisely not full-year employment, this rule was a systematic exclusion mechanism dressed in actuarial language. The industrial Lowlands absorbed the Beveridge model and, when those industries collapsed from the 1960s onward, lost not just their incomes but their structural eligibility for the contributory benefits the model had promised.
Connection Forward
Chapter 4 asks what was simultaneously present and defeated in the same decades: a materially specific, household-grounded counter-mathematics to the actuarial model, built from grocery receipts, welfare rights campaigns, and deprivation surveys. Understanding why that alternative lost requires understanding what the postwar welfare state had already built and why its administrative logic so consistently resolved toward simplification.