Open Source Government: Re-Imagining the Public Sector as a "Smart Grid"

Tuesday, October 11, 2011

Once upon a time, there were two nations, each vast and powerful. With its wealth one built a mighty railway, connecting its cities with ultra-modern trains where thousands would be whisked at hundreds of miles per hour. The other used its wealth to build a labyrinth of roads for millions of cars to carry its people wherever they wanted to go, at will, one or two at a time.

There came a day when the nation of cars realized the nation of trains could move more of its people for less fuel. With energy running short, the nation of cars came to envy the nation of trains, to regret its own decisions, and so it stopped building roads and started laying tracks. Just barely spared the fate of unsustainabilty, the reborn nation of also-trains lived happily ever after.

So runs the parable some sustainability advocates tell about American versus Chinese transportation policies. It's a compelling narrative. It is also deeply flawed.

Recapping the old debate: lonely bowlers vs. Big Brother

The progressive paradigm of the twentieth century sought to bring essential common services under public, that is to say central governmental, control, and to do so on a grand scale. This included everything from the uncontroversial, like municipal water utilities, to the contentious, like public schools and, of course, transit. The contentious systems, constantly pulled between those favoring public administration and those favoring privatization, have suffered for the lack of a consistent, coherent approach.

Big centralized alternatives planned, built, and administered by government -- like most social and industrial infrastructure is in Communist-Party China -- would likely be more efficient to operate than the uneven patchwork systems we have in the U.S. today. It is a familiar convention, left over from the debates of the Cold War, to ascribe that presumed efficiency to the virtue of civic collectivism as opposed to selfish individualism. Of course, it is equally conventional to accuse industrial socialism of producing unwieldy state behemoths that lack soul or sensitivity to each citizen's individual concerns, spirit, or creative insight.

Which type of system is more efficient?

For a moment, let's leave moral philosophy (and the Cold War) aside and ask, what kind of organization produces the most efficient system? Effenciency is not the only possible value, but it is a critical metric in the face of peak oil and climate change. One Block Off the Grid recently produced this infographic comparing a range of transportation modes by per-passenger efficiency.

And the winner is... unclear. The least-efficient modes are indeed private motor vehicles, your usual progressive planner's bogeymen. The most-efficient modes, however, are not that planner's darling trains. (Full disclosure: I also love trains. I love them beyond mere reason.) Trains do well, but are out-performed by buses and carpools -- creatures of much less-centralized systems. And then there are the zero-emissions vehicles: solar-powered electric cars and bicycles, arguably the modes most independent of any centralized system.

Of course, this infographic compares individual trips, not complete transportation systems. A fleet of buses requires roads, which bear energy costs of their own. An electric car fleet requires the manufacture and maintenance of all those cars. Trains require track. Physical infrastructure is a significant factor in the overall cost and energy efficiency of any possible transportation system.

That said, let's assume that manufacturing and construction are held to reasonable standards for any system. (Not currently the case for China's railways, alas.) Let's assume further that each system's phsyical infrastructure is used to near capacity -- that is, seats are generally filled. These assumptions should reduce the impact of physical infrastructure on comparitive efficiency, so that we can more or less leave that variable aside. They would eliminate single-occupancy cars (not carpools) from consideration altogether. And they would also introduce a problem for the systems with infrastructure requiring highly centralized administration, namely freeways and trains.

A shortcoming of the Progressive Era model

The problem is that it is difficult for central planning to anticipate future patterns of demand, and it is inefficient to build roads or track in what proves to be the wrong place. From a systems perspective, this is the greatest failure of the American freeways: roads were built in anticipation of demand, often in an attempt to force economic development along new thoroughfares. As Joel Kotkin and others have observed, American society is now burdened with the considerable opportunity cost of having thus built a lot of transportation, housing, and related infrastructure in places whose economic productivity going forward doesn't justify the initial investment.

This unsustainable "sprawl" is often regarded, at least by Americans, as a car problem. Though that may be a fair association in light of our experience these past sixty years, overbuilding is not a liability of cars (which were around before the sprawl) but rather of any mega project based on hoped-for, not present, demand.

Cars don't make sprawl. Bad central planning does.

Like any new technology, automobiles made many things possible, some good and some bad. It was giant acts of the public sector -- zoning vast swaths of land exclusively for low-density housing, and building freeways for its residents to access urban resources -- that re-fashioned American communities after the private car's darker temptations. We could have incorporated cars into the landscape in different ways, that would not have produced unsustainable exurbia.

Furthermore, I note a tendency -- in myself as well as other critics of freeway-based transportation -- to lump in sprawl's terrible opportunity cost with other complaints and level it all against "car culture," from its tendency to produce social stratification to the bland aesthetics of suburban streetscapes, as though these problems arise from cars themselves and not from public policy. It would be convenient, of course, for cars to be the consummate enemy, the single, unified source of all bad things; but with facile thinking comes a blindness to potential opportunities for real progress.

How about a good fleet of cars?

Just as Jaime Lerner exploded the assumption that rail-based transit outperforms buses, new ways of deploying smaller vehicles may be poised to challenge the primacy of fixed-route models in transportation planning. Of course, carpools aren't new, but they have traditionally been treated by planners as a way to make a disease -- people driving private cars from beyond mass transit's reach -- merely less virulent. Now, by contrast, smaller private vehicles are beginning to receive systemic coordination rather than being left to individuals to organize for themselves.

Some employers, such as Microsoft, have begun operating private bus services connecting their offices with residential concentrations of workers. So-called dollar vans in New York City's outer boroughs are picking up passengers bound for popular destinations who would not be willing to wait longer for the bus. And as the likes of ZipCar are confronted with the peer-to-peer approach of RelayRides, which with Detroit's cooperation could essentially create city-wide car sharing networks with no middle men, the trend appears to be toward increasingly decentralized systems of mass transit. What is especially notable is that, by bringing rides and riders closer together, this decentralizing process is also increasing the transportation system's overall rate of capacity usage -- which is just the opposite of what happened with the suburban decentralization.

Systems versus non-systems

In terms of fuel and overall cost efficiency, the type of vehicle used by any transportation system is trivial in comparison to the ratio of butts to seats. An SUV carrying five passengers is more efficient than a compact hybrid (to say nothing of a city bus) carrying only one. The conventional association of private vehicles with low and of public vehicles with high efficiency is in large part due to private vehicles generally not having been deployed in any kind of coordinated, systematic way. But what if they were? Now at last we can compare one system against another.

(Wrong! If coordinated such that each carries four people, it takes only 12 of those cars, one fifth of the 60 shown -- and their routing might prove more efficient than the bus's.)

Fixed-route systems have the advantage of economies of scale. They can be engineered to maximize the efficiency of the most-traveled routes. On the other hand, flexible-route systems have the advantage of adaptability and lower capital costs. They can bear trial and error, and they aren't expected to satisfy all possible stakeholders.

Why not, then, have our cake and eat it too, deploying fixed routes where demand is already established, and faciliting a more agile system to fill in the unknowns? It seems the two could be quite symbiotic: robust, centrally-operated trunk lines provide convenient points for gathering and dispersing the decentralized system's resources; and travel patterns that emerge in a flexible-route system inform future investment in fixed-route infrastructure.

Now, isn't that just a commuter rail line with park and ride lots? (Which, I've been delighted to discover, DC Metro calls the "Kiss and Ride.") In a word, no. A commuter rail line is a centrally-administered trunk system, but what happens when drivers leave the park and ride is not systematically coordinated, and consequently tends not to maximize the use of vehicle capacity. It's inefficient because it's disorganized. A commuter rail line interfacing primarily with shuttle vans, mobile-app-facilitated carpools, and neighborhood bicycle routes would look substantially different, and would require less energy per rider. 

Non-centralized public works and "open source"

This begs the question: what is the public sector's role in facilitating decentralized systems? Today's political conservative would probably answer, "by getting out of the way," but this approach has not proven to work very well (at least not for a majority of the public).

A better model can be found in the world of open-source computer software, in which it is standard practice for a "core" team to maintain a common set of functions, those that most users require, and to provide "APIs" for other programmers, to facilitate their development of relatively niche functions that build off of the core functions. In mature, successful open-source software projects, often the core team's work will consist as much in maintaining (and documenting) tools and methodologies for other programmers to implement, as it consists in making free-standing functionality that one would use on its own.

Analogously, an "open source" public sector would understand its role not necessarily as maintaining central systems to serve all citizens, but more as providing regulation, coordination, and common resources to facilitate myriad local systems.

This isn't just about cars and trains

Americans tend, rightly, to take pride in our water utilities. A system that reliably supplies potable water to a quarter of a million people is no small accomplishment. However, as Robert Glennon and others warn, the system's current course, given population growth and climate change, is unsustainable. Glennon spoke in Seattle last spring and described Patricia Mulroy's valiant efforts to keep water flowing in Las Vegas. Counterintuitively, the (in)famous casinos of the Las Vegas Strip, with their massive hotels and pools and fountains, are now part of the solution, not the problem. This is significantly due to Mulroy's use of the regional water utility to collaboratively re-engineer the casinos as miniature water utilities in their own right. (One casino owner, Stephen Wynn, calls her "the best public servant I've met in my 40 years on the Strip," because her Southern Nevada Water Authority showed his business how to save money. How? By doing what the water utility couldn't do itself, which is to implement efficient water flows inside the casino's operation.)

Similarly, DC's water utility just got the EPA to take one big step toward recognizing the public water system as more than just pipes, pumps, and treatment plants. DC Water's general manager George Hawkins spoke last month at the Greening Greater Washington seminar about two possible solutions to the city's severe storm runoff problem: one giant tunnel, or a thousand green roofs. The tunnel mega-project is classic Progressive Era-style public works, and it's approved and ready to go, but Hawkins would rather use his resources to transform infrastructure that DC Water doesn't own. Nurturing a "spongier" city of buildings and grounds would effectively disperse much of the water utility's function into the natural, and largely passive, function of the whole built environment. This constitues a vision for a new kind of public utility, one which Hawkins (who ought to know) believes would be more efficient and less expensive.

Better than mere efficiency, resilience

As I wrote last week, "it is risky to have essential systems, upon which all local communities depend, operated by monopolies -- because monopolies are prone to stagnate, charge in the wrong direction, ignore local customers, and be co-opted by money- or power-seeking individuals." This is not a case against either the private or the public sector. It is a case against giantism, against that massive drainage tunnel, and in favor of the thousand green roofs. Any personal investment advisor will encourage a diverse portfolio, to insure against  catastrophic loss if one or two stocks fail. Why should investment in society's commons be any different?

Internet access, now recognized as a human right, is still controlled mainly by big corporations (as in the U.S.) or by big government (as in China), and neither arrangement has fulfilled the ideal of universal access. Meanwhile, my neighborhood in DC, Bloomingdale, is covered by a community mesh network, and though I pay for commercial internet access for my work, when it goes down the neighborhood's network is always there.

Municipal government efforts to provide city-wide internet access have generally not fared well, typically for financial reasons, but it seems public facilitation of neighborhood mesh networks would be considerably less costly. Then, once many contiguous neighborhoods' networks were established, connecting them to each other would be natural and easy, if not automatic, and would result in a system both inexpensive and resilient.

This is not an anarchic, libertarian proposition. It envisages an essential new role for public administration in facilitating and regulating dispersed systems, in addition to its well-established role maintaining central trunk resources.

The smart grid evokes electrical power, and open source evokes digital software or the likes of Wikipedia -- but the same hybrid central-distributed approach described by both deserves consideration for all aspects of the commons. It should be possible, with a little creativity, to re-imagine any social system with the central administration of core resources and methods working in concert with distributed administration of services for local communities by local communities.

So, what would a "smart grid" look like for food? How might an "open source" education system work? I figure we'd might as well start brainstorming now, because this is probably our future.

%s1 / %s2

Click here to leave a comment the old-fashioned way, without using Facebook