A memorable turn of phrase in Charles Darwin's correspondence from the 1850s distinguished two tendencies among contemporary taxonomists, themselves labelled either ‘lumpers’ or ‘splitters’. Faced with the mind-boggling variety of life on display and charged with picking out salient features for its categorization, naturalists could emphasize either the broad similarities shared across, or the fine distinctions between, a given set of phenomena. Dan Bouk's ambitious and lively history of US insurance explores the array of tensions which arose during the parallel process of parsing, recombining and interpreting vital statistics within the human science of life assurance, as it was practised around the turn of the twentieth century by those in the risky and profitable business of guessing human longevity.

Students of the history of knowledge should not be surprised that this process of enumerating and quantifying was shaped by the social, political and economic context within which it occurred and, thus, the ‘correctness’ of these numbers has been the source of great controversy. The burgeoning literature on insurance – with its interlocking debts to the histories of probability, statistics, medicine and business, and on which Bouk draws extensively in his copiously researched study – means that its importance as a topic for historians of the sciences is now firmly established. What Bouk's study adds to this body of work is not so much any single thesis about insurance in America as a fascinating account of the ways in which what he terms the ‘by-products’ of insurance rippled outwards to influence other spheres, from social security to the diffusion of weighing scales and the ritual of the annual medical checkup. Thanks to the path-breaking work of Timothy Alborn, readers may be familiar with the ways British insurers responded to the political, cultural and economic pressures of the late nineteenth century, but Bouk's focus on the US insurance market has the virtue of drawing our attention to the specifically American flashpoint of race, which illuminates in new ways what has elsewhere been described as ‘insurantial’ thinking.

Through its commitment to pricing risk correctly, one might expect actuarial science to be, as it were, colour-blind; yet Bouk's account reveals the ways in which dubious racial categories – not only ‘black’ but also a monolithic idea of ‘white’ – conspired to make it more expensive for black Americans to contract life assurance than should otherwise have been the case. It was eventually in the courts and state legislatures that the fight against racial prejudice took place, in a series of moves which would outlaw discrimination – in a direct challenge to the splitting and lumping (technically, ‘classing’ and ‘smoothing’) fundamental to the insurance practice of computing risk. This particular flashpoint – between insurance and anti-discrimination – provides the main pivot of the book and raises questions about the very possibility of providing insurance equitably, based on the traditional pooling of risk, within a society which claims to be both democratic and capitalist. The fascination of insurance is thus its potential to raise these delicious (or rather unsavoury) ironies. Dominique Pestre has recently lamented the reluctance of scholars in STS to employ the more traditional tools of political analysis in their work, and so it is that when Bouk describes his work as a study of ‘cultural’ artefacts he undersells what is, in fact, a major contribution to thinking about the political question of insurance.

The book is structured around a somewhat discrete set of vignettes illustrating a more general phenomenon, after which each chapter is succinctly titled. ‘Classing’ recounts the scandal of the failed American Popular Life company, whose founder faced trial for misleading the public, but whose importance for Bouk's purposes lies more in his belief in splitting incessantly – employing a novel phrenological pseudoscience to produce the most individually tailored risks possible. ‘Fatalizing’ explores the beginnings of resistance (prompted by disgruntled black consumers) to insurers’ use of population statistics as the basis for their ‘risk-making’. ‘Writing’ cleverly explores the manner in which doctors and medical investigators coalesced into Dastonian ‘observational communities’, while ‘Smoothing’ – perhaps the strongest chapter because of the detail it provides into the role played by an actuary, the legendary Emory McClintock – builds on the work of Will Ashworth in stressing the continuities between the shared astronomical and actuarial practice of smoothing data, a practice which, while perfectly normal to a ‘risk-maker’, confounded and outraged the public. Chapter 5, ‘A modern conception of death’, changes tack with the tale of Frederick Hoffman – a social reformer who found a home within the insurance industry – tramping around southern graveyards totting up lifespans to make good the lack of reliable data on those communities. Chapter 6, ‘Valuing lives, in four movements’, looks at the lesser-known vocation of the modernist composer Charles Ives as a New York insurance agent and the problem of valuing human lives, especially those of women. Chapter 7, ‘Failing the future’, explores what Bouk terms ‘white data politics’, or the statistical parsing of the white immigrant community in an atmosphere of eugenics, a story whose payoff makes sense of how actuaries came to price black lives more generally.

One frustration upon reading this rich and provocative book is Bouk's reluctance to evaluate the practices of classing and smoothing he identifies as the signal features of insurance. The political and ethical complexity of these issues certainly militates against a simplistic judgement of insurers’ practice, and yet one senses – reading between the lines – an implicit disapproval of his ‘risk-makers’, that the insurers are the baddies of the story. And yet one could probably argue, on the evidence presented, that insurers behaved perfectly rationally, equitably and fairly, given the data they possessed: discriminators will discriminate. Such moral questions are shown to recur into the present in an epilogue which brings the story rapidly forward into the age of big data and telematics, phenomena which imbue the history with added pertinence. This book contains such a wealth of material and such a dizzying cast list that some readers may feel overwhelmed by the thickness of the description, nonetheless it is written in fluent and breezy prose, and full of playful intent. The snappy chapter titles, however, are a distraction – from the misleading to the perplexing – and do a disservice to the quality of the material, which might have been better organized in a more user-friendly manner. Despite such minor quibbles, here is a substantial book of huge value to readers interested in the politics and the business of statistics, medicine and race.