The Enumerative Institution
Robert Gerard Pietrusko
Census enumeration, land use analysis, the measurement of labor inputs, and other instances of large-scale quantification are the preferred representational forms for bureaucrats, economists, developers, and politicians. Through highly-refined classification methods, unruly differences within a population are aggregated into discrete categories with parameterized attributes—the average, the median, the rate per 1000, the standard deviation. Such analytical tools are believed to reproduce everyday life and its social organization in a new and manageable form. The tabulated spreadsheet is the preferred map of speculation for numerous agencies, where the space of lived experience can be structured into an ordered grid, easily shaped by the current stakeholders and their varied agendas. Despite the claims of an increasingly precise demographic technique, these tools are haunted by a looming obscurity in their calculations; every mathematical aggregate creates a gap between representation (the “maximum likelihood estimation”) and the underlying reality it purports to represent—a gap commonly known as the Statistical Cover.
In 2012, The Enumerative Institution was established by the New York State Regional Council to spearhead a multi-year research project on the social uses of the Statistical Cover. Though long assumed to be a mathematical formalism with no real relevance to the fields of demography and public policy, it had recently become a topic of official interest after the publication of three critical reports by the Government Accountability Office (GAO-11-154, GAO-10-452T, and GAO-10-430T). Across roughly seventy pages, the reports highlighted major mathematical issues in the enumeration techniques deployed by the U.S. Census Bureau, including the acknowledgement of the Statistical Cover itself. Specifically, they stated that beneath every statistical distribution is the condition of possibility for an infinite number of socio-economic systems, all of which would appear mathematically identical, once tabulated. Analysts speculating at the level of the spreadsheet could no longer make claims about the actual activities of the population it represented. Though the numbers were assumed to depict rather banal patterns of daily living and business as usual, one could not be mathematically certain that something radically different (yet statistically identical) was not being measured instead. More startling than the questions of accuracy, however, was the latent implication that the Cover could be purposefully exploited by citizens to hide alternative socio-economic arrangements. The GAO termed this potential phenomenon a “differential account,” and made specific recommendations for further research, ultimately concluding:
Moving forward, it will be important for the Bureau to identify how rapid changes in [demographic] technology and the public’s use of them may affect the effectiveness of its ef for ts to improve census accuracy, both overall and in terms of reducing dif ferential accounts.
(GAO-11-154, p. 13)
The Enumerative Institution’s initial findings confirmed, at least theoretically, that beneath the cover of statistical representation alternative societies were in fact possible and that the socio-spatial organization of these societies would appear mathematically identical to the more standard patterns of everyday life. Furthermore, this was found to be the case in every domain the census enumeration measures—the relationship between families and the sexes, the rate of population reproduction, the temporal structure of labor, the use of technology, the functioning of private space, the circulation of use-values and currency, and all other such metrics. On the surface, these findings appeared to discredit the specific techniques used by the Census Bureau, and the representational validity of large-scale quantitative measurements in general. The Institution, however, offered a different interpretation of their report. They highlighted the radical potential latent within the statistical cover. It showed how demographers assumed populations were not affected by—and actively responding to—their statistical representation, when in truth citizens were increasingly engaged with these categories and were finding new ways to repurpose them. The Census Bureau needed to account for the fact that they were providing people the raw material for producing their identities and understanding their own daily lives. Secondly, the Institution believed that the Cover created socio-spatial black boxes where an infinite number of experimental communities could be designed, enacted, and ultimately compared. Such studies would recover a foundational research agenda in the field of political economy—the original domain of demographic studies. They stated that “[an] empirical understanding of competing socio-economic models and the variety of forms collective life might assume when driven only by fundamental human desires” was conceivable.
The Enumerative Institution saw in itself the ideal organizational form through which these experiments could take place—namely, the legally-defined “institution.” With its managed organization of members, by-laws, material flows, relations to technology, and representational outputs, this formulation allowed society, city, community, and corporation, to all be confounded with each other. As a test case, it proposed that each of New York State’s 15,464 census block groups be treated as 15,464 such institutions—each with the goal of reproducing everyday life using an alternative socioeconomic structure.
In terms of practical methodology, the report again returned to disciplinary roots:
Seigneur de Vauban, engineer and advisor to Louis XIV, lamented in 1772 that one cannot know the city—the precise number of subjects; the true state of their wealth or poverty; what they do; how they live; their commerce and employment; if they are ill or well. The frustration of [demography] has been the same since. Though commonly understood as economically motivated, Vauban states that a true understanding of how people live would be of primary interest, thus betraying a desire to measure the organizational strategy of everyday life. Vauban was calling for an officially sanctioned taxonomy of this everyday life. Though subverted, the goal is still active within our Census Bureaus today and should be of renewed experimental concern.
Informed by the long history of typological analysis in the social and administrative sciences, the Institution established such a taxonomy, with thirty independent components. Each experimental social organization would embody some subset of these thirty features. Using over two hundred demographic and environmental indicators (and a multi-criterion optimization routine) all 15,464 NY block groups could be ranked based on their statistical potential for hiding the proposed community beneath their cover.
In the Institute’s first empirical study, five sites were algorithmically chose —360470052011, 360610203001, 360810489002, 360470420001, 361031699021. Observing only the sites’ statistical representations, a speculative ethnography was conducted to imagine the social and material relations, spatial processes, and “structure of feeling” within each of their “black boxes.” The results were recorded through interviews, reportorial field notes, and site plans.
Conclusion
The liberating potential of statistical over-representation is not a struggle for legibility as so many proponents of “big data” claim, but rather, for new and productive strategies of obscurity. Likewise,
increasingly high-resolution measurements of the population and its resources do not result in an over-bearing surveillance system; instead, they generate an exhaustive pattern of statistical black
boxes across the American landscape. Within these black boxes, an infinite number of social and spatial experiments may take place— designed, enacted, and judged by the communities themselves. Our daily patterns and social relations may be freely determined while still aggregating, numerically, into the “Average.” To design this life, statistical representation becomes a new site constraint, a new boundary condition, and a potentially utopian closure within which radically new organizations of life may be modeled.