What is Urban Resilience?

It has become standard practice on this barely-managed blog for me to begin each post with an apology/ apologia for not having contributed lately, and I suppose this is no exception. I have been up to my neck in dissertation stuff, workshops, conferences, faculty meetings, etc. I would, however, like to make a habit of contributing small, more manageable snippets of things that I am thinking about from time to time.

To start it off, I'd like to talk about the concept of resilience in conflict studies. This topic has particular significance for me, as a partner in the nascent "Urban Resilience and Chronic Violence" group. During our initial discussions, the definition of resilience (or, in that case, urban resilience) took on predictably critical importance, but it was elusive. There are several questions that were brought up, among which four stand out:

1. Does resilience refer to the ability of a system to "absorb" shocks without fundamentally transforming or adapting, or does it refer specifically to the ability to adapt?
2. In a potentially related question, does resilience refer to how much change or turbulence a system can absorb, or how fast it can rebound after the shock?
3. Do all urban "sub-systems" necessarily exhibit resilience in mutually beneficial ways, are do certain forms of resilience compromise others'?
4. How does urban resilience in conflict affect the capacity and legitimacy of the state?

Conflict studies has borrowed heavily from environmental science in attempting to define resilience in human societies. It turns out that the second question has already led to two different definitions of ecological resilience among environmental studies folks, as shown in Figure 1:

Figure 1: Two forms of resilience. From Adger, Neil W. (2000). Social and Ecological Resilience: Are They Related? [Thanks to Ami Carpenter of the Kroc School of Peace Studies for introducing me to this.]In the first conception, there is a threshold of "wellbeing" or "output" or something (on the Y-axis) below which any system will be unable to recover from a shock. The greater the resilience of any given community, then, the closer to zero that threshold will be, such that a shock of any magnitude will still be followed by positive growth to reestablish a status quo.

Problem 1
My first concern with this conception is that it confounds the "break" and "sustain" points. That is, it implies that the threshold would operate in the same way whether you were approaching it from above or (though this isn't specifically shown in the figure) from below. The former would be the case if the system was being adversely impacted, while the latter would be the case in a post-crisis recovery mode, when the system was looking to attain a former high performance level (as in the right-hand side of figure 1, after the initial shock).

In fact, the two aren't necessarily the same. To see this, we might take a well known economic model of economic growth - the Keynesian model used by Pred to demonstrate "cumulative causation" in capital accumulation. Krugman, Fujita and Venables take the model, give it a basic dynamism, and then solve for its equilibria.

If we say that output, Y, is a function of some exogenous demand, X, multiplied by the inverse of 1-a, where a is the proportion of the knock-on demand resulting from the initial impetus that is to be met internal to the system in question (here, taken to be the city), then we can write:

1. Y = [1/(1-a)] X

The way we have set this up, we can now introduce some dynamics into the model, following the lead of Fujita, Krugman and Venables, positing that a itself, up to a point, will be a function of how much social capital was created in the last round, up to some maximum (say, the maximum number of interactions that can be formed in a group given the number of individuals):

2. a(t) = min [aY(t-1), a(bar)]

We can then rewrite (1) such that

3. Y = X/(1-aY')

with equilibria at

4. Y = (1 +/- root[1 - 4aX]) / 2a

for a < a(bar), and

Y = X/(1-a)

for a >= a(bar).

Again, following Fujita, Krugman and Venables, we then have a function that looks like the following, given a high enough a.

Figure 2. Output as a function of exogenous economic demand in the Pred cumulative causation model (from Fujita, Krugman and Vanables).

Note that there is an overlap of these two functions such that radically different outcomes are associated with the same exogenous demand depending on whether the system is suffering decline from a high performance level (i.e., moving from right to left) or recovering from a crisis that dropped output precipitously from a(bar)/a down to the lower level (i.e., moving from left to right).

Of course, the same sort of argument could be applied to social capital, with X possibly representing some underlying economic driver, with higher levels of economic demand generally yielding more social interactions through production networks or something. Some accommodation would have to be made for the distinction between social capital deepening versus widening (just as in the Solow-Swann model). That is, if we assume that all social interactions are precipitated by some sort of exogenous economic demand, the model still doesn't specify as to whether the resulting interactions are "new" or simply repeated activations of "old" relationships. At this point, the model would have to rely greatly on sociological applications of network theory, in which the morphology of the networked society - for instance, polarized, hierarchical, multi-polar, etc. - play a role in the probability of any one interaction creating "new" social capital.

Leaving all that aside, though, we can still postulate that when the driver of social relationships (be it economic demand or something else) drops from a high level, the "break point" for the society will be found at a lower level than the "sustain point" (or "recovery point"). In the model, above the "sustain point" and below the "break point," both declining and a recovering economies have the same social capital at comparable levels of economic activity. Between the two, however, the declining economy will have higher levels of social capital - and once it dips below the "sustain point," the economy will have to grow quite a lot more to achieve comparable levels of social capital as it had before. In this scenario, it's easier to salvage social capital than it is to build it - and therefore there is a strong argument for preventative measures to forestall economic and social decline in the run-up to violent conflict, rather than applying the proverbial pound of cure.

Problem 2

Arguably, both of the forms of resilience rely upon redundancy. That is, systems that have redundant pathways, organs, functionality, etc., tend to be more resilient to shocks. For instance, the human body has redundant kidneys, eyes, ears, hands, etc. And while having two of each thing has distinct advantages in terms of utility in the here and now (e.g., depth perception, aural directionality, etc.), it also conveys advantages in the event of disaster (if one hand is bitten off by a tiger, the other one is still around, etc.). Ecological systems operate similarly, such that the greater redundancy there is in terms of ecological niches, the less systemically damaging it will be for any particular species to decline in number.

Fine. Point taken. But humans societies are different in one critical way from other ecological systems: its constituents convey complex information on possible ways to structure organizations to one another by way of language, rather than simply via DNA by way of natural selection. That is, humans are strategic animals who have a (possibly limited) capacity to plan for the future. Resilience in human societies - and, by extension, in urban areas - then may not simply be judged as the ability to absorb or respond post facto to a crisis, but also to recognize the signs of a future crisis and envision (and I stress that this is a creative, as much as analytical, process) series of contingency plans.

This creative impulse is often ignored by academics, including myself. Just last week at my doctoral dissertation colloquium, I made an analogy between production firms and the rebel groups that prey on their production networks, likening their relationship to that between a population and a disease, in which over the course of a series of attacks (or outbreaks), the "immune system" of the firms is strengthened by virtue of supply chain adaptations, while the "virulence" of the rebels also decreases so as not to kill the golden egg-laying goose. Roger Petersen, one of my committee members, made the point that firms and their managers have the capacity to strategize; they don't simply happen to have the type of production networks that lessen the effects of predation, but consciously choose to develop them.

And so?

Would it be too trite to sum up with some song and dance about how resilience is composed of preventative, peri-crisis, and post-crisis forms of redundancy? Probably. In the end, resiliency seems like a word that conflates a number of different ideas that are generally considered to be "good." And yet, as Diane Davis at MIT has repeatedly reminded me, resiliency may generate new challenges. She cites the growth of private security forces in many developing countries: they help to fill a vacuum left by the state in property rights enforcement, but they also ultimately undermine the Weberian legitimacy of the state as the monopoly wielder of violence.

I would add that the same might be said of urban industries and economic governance. These firms' production networks become so splintered and reticulated (and de facto localized) that it becomes difficult for a weak state to tax their purchases. If all inputs originally went through a single international port that was easily controlled by revenue officials, the task becomes much more difficult in the event of rebel takeover of the port, when inputs are now being redirected across porous borders or bought from domestic producers and traders.

Economic resiliency in this case is at odds not with the state's monopoly on the use of violence, per se, but with an intimately related would-be (should-be?) monopoly on taxation. Charles Tilly saw these two monopolies as intertwined in state-formation: the monopoly on violence meant that no other actor could force citizens to pay for public goods. State formation then grew out of a sort of protection racket in which citizens' voice and power grew in importance, demanding more and more concessions of public goods provisions on the part of the state.

All of this raises the question of postwar recovery for me. Weak states wishing to regain their twin monopolies will actually have an incentive to ally themselves with international capital and commodities markets, because imported inputs are easily taxed. But the sheltered economies that developed in wartime came to rely on much more redundant and resilient networks of inputs - and ones that thus linked urban areas with rural areas in more mutually beneficial ways, despite the drastic economic decline that war obviously precipitates.

Resilience in this environment pits two very different visions of nation against one another: the nation conceived as economically cohering process, and that conceived as state-legitimating process. It might be posited that not until transportation costs fell in an era of vast production capacity disparities did the two ever come into direct conflict... Wow, that grew into something more than small and manageable.