Anthony M. Townsend - Smart Cities - Big Data, Civic Hackers, and the Quest for a New Utopia

Cybernetics provided a theoretical wrapper for the more mundane field of operations research, which also grew out of wartime planning and applied the new science of systems to the simulation and planning of large organizations. These ideas were deeply embedded in the design of massive, networked organizations like the air defense system coordinated by SAGE. But it didn’t take long for cyberneticians to turn these techniques, and the new power of computers, to the problem of America’s cities. Like Seldon, they made hasty approximations as they rushed to twist a complex urban reality into a computable set of equations. But unlike the psychohistorians in Foundation, whose doomsday prophecies were fulfilled by the story’s end, the real world cyberneticians never succeeded in building a machine that could predict the city. In fact, they failed. And that failure had terrible consequences.

As a grad student at MIT in the late 1990s, one wintry afternoon in the library I stumbled upon a curious book called Urban Dynamics, by Jay Forrester. I was spellbound to discover in its musty pages an entire science of cities, seemingly forgotten for decades, laid out in objective prose and logical cybernetic flowcharts. Like Wiener, Forrester was a professor at MIT and had also worked on targeting systems during the war. But his subsequent interest in cybernetics was more practical. During the 1950s, Forrester co-led the design of SAGE, a masterstroke of cybernetics that linked up dozens of control bunkers with over one hundred radar stations throughout North America.

Forrester’s experience building SAGE taught him that engineering wasn’t the biggest obstacle to building big, complex technical systems.

The real challenge lay in managing the people and organizations who would use them. Humans, it turned out, were far harder to understand and control than machines.50 Beginning in 1956 at MIT’s new Sloan School of Management, he quickly became one of the leading lights in operations research. While cyberneticians like Wiener debated the nature of the universe elsewhere on campus, Forrester was more interested in actually designing really complex things. He developed techniques for mathematically modeling industrial systems, focusing on how feedback loops and time delays governed flows and stockpiles of resources and products. The culmination of that work, Industrial Dynamics , was published in 1961. It analyzed the workings of a General Electric plant in Kentucky, laying the foundations for modern supply-chain management.

Having mastered the corporation, Forrester looked for other complex systems to which he could apply the cybernetic tool kit he now called generally “system dynamics.” When former Boston mayor John Collins was appointed as a visiting professor of urban affairs at MIT and, by sheer coincidence, moved into the office next door, Forrester seized the opportunity.

Forrester wasn’t the first to get the idea that computer models could be used to understand cities. The success of systems engineering in the massively complex defense and aerospace sectors held out the hope that it was up to the task of city management. It was a time of great anxiety about the future of American cities. Summertime riots had become an almost annual event in inner cities, as jobs and the well-to-do fled for new suburbs. As Forrester wrote in the introduction to Urban Dynamics , published in 1969, “The plight of our older cities is today the social problem of greatest domestic visibility and public concern.” Using Collins’s connections, he canvassed experts on a range of urban issues. He developed equations that described how various parts of the city operated—housing and labor markets, for instance—and how they interacted with each other. These relationships were programmed into a computer to create a simulation that purported to explain how cities grow, stagnate, decline, and recover.

Rather than studying a particular city, Urban Dynamics was an attempt to abstract a generic system model of cities. But the book confounded urban policy makers, not just because of its lack of grounding in an actual place but because of its counterintuitive conclusions. Forrester’s generic city started in a “stagnant condition” that seemed to characterize most big US cities at the time—a stable equilibrium of high unemployment, a surplus of slum housing, and a shortage of housing for professionals. But setting the model to simulate prevailing urban policy, such as job training for the unemployed and direct federal aid to cities, actually resulted in worse outcomes. Even more surprisingly, the results argued in favor of the policy of demolishing slums and replacing them with high-end commercial and residential buildings, a tactic that by the end of the 1960s was already highly controversial. Nevertheless, Forrester had an “unflinching confidence” in his methods and the results, as a book reviewer in the Journal of the American Institute of Planners put it.54 He casually excused the book’s lack of reference to any contemporary work in urban studies. “There are indeed relevant studies on urban behavior and urban dynamics,” he wrote, “but to identify these is a large and separate task.” With no formal training in urban planning, based solely on his computer simulation, Forrester recommended the demolition not only of slums but of federally subsidized public housing as well, which the model showed became poverty traps for their inhabitants. While the ghettoization of the poor in housing projects is now widely recognized, it was obvious failures like the disastrous Pruitt-Igoe complex in Saint Louis (which was torn down in the 1970s) and painstaking fieldwork by a generation of social scientists that made the case in the end.

Urban Dynamics was perhaps the most ambitious effort of that generation of computer-based urban simulations. But it came at the tail end of a decade of failures to apply systems analysis to urban problems. As historian and sociologist Jennifer Light explains in From Warfare to Welfare: Defense Intellectuals and Urban Problems in Cold War America, much as IBM turned to cities for new business during the 2008 financial crisis, the defense industry began looking for new markets for military computer technologies almost as soon as they were invented. As early as 1957, connections were being drawn between the similarities of military planning and urban planning.56 Uncertain how long the Cold War would sustain defense spending, Light argues, the think tanks “decided that the survivability of their organizations depended on finding ways to transfer their innovations beyond military markets.” In the late 1950s, defense contractors such as TRW and RAND began publishing studies in urban and public administration journals, Light recounts, “suggesting how techniques and technologies from military operations research such as systems analysis and computer simulations might offer a new direction for city management.”

The results were less than impressive. In the early 1960s, as part of its federally funded Community Renewal Program, the city of Pittsburgh attempted to develop computer simulations that would forecast the impacts of public spending decisions about transportation, land use, and social services. Almost immediately, problems appeared. One program that sought to measure the impact of housing clearance for an expressway produced nonsensical results. Rather than expand the city’s capacity and inform better decisions, technology constrained thinking. As Light explains, Pittsburgh’s planners “realized they were shaping their questions and problems to fit what could be modeled ... yet rather than characterize this as a flaw of simulation techniques, they used this finding to justify why one would want to use them.” Captured by the computers’ limits, they argued that simpler models were better. In their words, complex models that were “photographic reproductions of reality ... would be so complicated that they would be of little, if any use.”59 With nothing usable to show for its modeling efforts, in 1964 the city fired the projects director and declined to apply for an extension of its federal funding for the effort.