Achieving societal goals requires understanding how action leads to consequence. The recent Colorado fire disaster shows the complexities of causality.
Before jumping into this discussion, I’d ask that you considering donating to the Marshall fire relief efforts — the Community Foundation of Boulder is one good option.
In 2022, I intend to write more about policy research and its connections to policy practice. In this post I discuss a specific conception of causality, which I am calling policy causality. Goal-seeking collective action would be impossible without the ability to anticipate the connection of an action with a consequence, with some degree of reliability. Policy causality thus reflects our expectations for the relationship of decisions and outcomes. Such expectations can be reliable or they can be flawed — obviously, we’d prefer them to be reliable. Policy research, which is conducted by governments, academia, journalists, non-profits and others, often focuses on improving our understandings of policy causality in hopes of improving decision making.
In this post I consider the fire disaster which occurred last week in my community, called the Marshall fire. The fire destroyed almost 1,000 homes and has left 2 people missing. To avoid similar future disasters, it is important to understand (a) why the Marshall fire disaster occurred and (b) what steps might be taken to make sure that a repeat does not happen again. The former focus often is called policy evaluation and the latter, policy implementation. Evaluation and implementation are obviously related through a connection of lessons learned and actions taken.
A first key point to make absolutely explicit: The desired outcome of collective action must be made perfectly clear if explorations of policy evaluation and implementation are to be most useful. In this instance, the outcome to be avoided, to the extent possible or practical, is a destructive and deadly fire disaster. That is not necessarily the same thing as avoiding fire, the presence of which may actually be important to reducing risks of disasters, such as through controlled burns to lessen fuel loads. So asking “why did the fire disaster occur?” is a very different question than “why did the fire occur?” Such precision in problem formulation is central to effective policy inquiry.
Understanding policy causality in the context of disaster risk reduction thus requires answering a few key questions. Among them:
Why did the Marshall fire disaster happen?
What actions, in the context of the tapestry of existing policies, can reduce the risks of future fire disasters?
Right away we find ourselves in the midst of questions of causality, both retrospective and prospective. Understanding policy causality requires an understanding of several subsidiary forms of causality, which are important for understanding policy casualty, but which are not the same thing.
These subsidiary forms of causality are:
Proximal causality. This refers to the specific chain of events which led to the disaster. Among the potential factors of proximal causality in the Marshall fire disaster: an ignition source (to be determined, but without a doubt human caused), an extreme wind storm, and fuel for the fire to spread (in this case, dry grasslands resulting from a wet spring followed by an extremely dry summer and fall). The image above shows the grassland that burned prior to the fire reaching the communities of Superior and Louisville. Additional factors of proximal causality include exposed flammable structures, capabilities (or not) to halt an advancing fire in 100 mph winds, as well as warnings and evacuation procedures. This is not meant to be an exhaustive list, but indicative of the specific causal chain of events last week that led to the disaster.
Distal causality. This refers to the underlying or background conditions that may have set the stage for the fire. Among the potential factors of distal causality: regulations and enforcement of open burning and decisions in the preceding months and years about prescribed burns in the public open space neighboring the communities. More fundamental decisions that set the stage for the disaster go back years and even decades, such as the initial decisions to create large areas of public open space on the edge of tightly built neighborhoods. And of course, considerations of human-caused climate change (and of course climate variability) as a factor in loading the fuel source (either via a wet spring that led to grass growth or the subsequent dry summer and fall which dried it out). The figure below shows a time series of drought in Boulder, CO for 2000 to 2021, via the US Drought Monitor.
Elements of proximal and distal causality can be more or less relevant to policy causality. Key questions to ask when considering the relevance of proximal and distal causality for policy causality include:
What factors were necessary and/or sufficient for the disaster to occur? For instance, without the ignition, the disaster would have not occurred. Similarly, without the extreme winds, the disaster would not have occurred. However, it is unclear (at least to me) what role prescribed burns or a less wet spring/less dry summer and fall might have played in risk reduction.
What factors are amenable to policy actions? Following the examples immediately above: Ignition can, to some degree, be controlled as it is the result of human choices, whereas extreme winds cannot be controlled. Thus, policy implementation focused on reducing or eliminating controlled illegal open burning has much greater chance for success than would policies focused on stopping the wind. Similarly, policies focused on risk reduction via controlled burns or altering seasonal precipitation patterns would need to be evaluated for their potential efficacy in actual risk reduction on specific time scales.
A goal of understanding policy causality, both retrospective and prospective, is to understand what possible actions might be taken that lead to a reduction of future risk, and not in theory but in practice.
Two other forms of causality can be useful, but also can mislead if we are not careful.
Model causality. In research conducted to inform policy, it is common to create simple or sophisticated computer simulations that are meant to reflect the real world. Such models are potentially very useful because they allow us to test policy implementation alternatives in order to assess their potential effects. Such model tests can thus inform both policy evaluation and implementation. But model tests can also mislead, if for no other reason than the fact that a model simulation of the world is not the real world. The use and misuse of simulations in establishing policy causality is complex and nuanced, and it is not uncommon for researchers or decisionmakers to confuse model inputs or outputs with the real world, and be led astray in resulting policy discussions.
Narrative causality. It is common in both research and in the media to establish causality by telling plausible stories of causality. For instance, if you were told that the Marshall fire started because some power lines blew down in the strong winds and ignited a grass fire, you would no doubt find that to be a plausible narrative that explains the proximal origins of the fire. Indeed, the claim of downed power lines were the initial narrative advanced to explain the proximal origins of the fire. But that narrative was wrong. Narratives are important means of communication, which also makes them useful in politics. Just because a story sounds plausible (or even appealing) does not make it an accurate description of causality or a useful guide to effective policy. The risk of course is that incorrect or misleading narratives support misplaced or ineffective polices. We are all familiar with dueling narratives on complex issues like policy responses to COVID-19, where it often seems that political gain has supplanted policy effectiveness.
So what are the implications of this extended discussion of policy causality? I can think of several.
First, whenever a goal is to improve future decisions, then it is important to ground policy evaluation (what happened?) and policy implementation (what might we do?) in evidence and argument. While models and narratives can inform judgments of policy causality, they are no substitute for rigorous understandings of proximal and distal causes of the outcomes that we have observed, and for projections of how those outcomes might be different in the context of alternative future decisions. Establishing policy causality typically requires technical expertise and the integration of different forms of knowledge — The case study of the Space Shuttle Challenge disaster offers an excellent example.
Second, when making a claim of policy causality, it is important to then associate that cause with possible actions to address that cause in the future. For instance, if a fire disaster resulted from a downed power lines that would imply a certain set of future actions. However, if the ignition source was illegal open burning, that would implicate an entirely different set of potential policy actions. Often distal causality factors do not lend themselves to immediate actions, as interesting as they may be. For instance, understanding the historical relationship of Boulder’s development in the 1960s and 1970s to its open space policies, and how that may have led to neighborhoods at fire risk in surrounding small communities is certainly fascinating, but that history might not have immediate policy relevance for reducing the ongoing risks faced by these communities, which now exist on the edge of those open spaces.
Third, understand that policy causality is typically complex, nuanced, often contested and even contradictory. Establishing policy causality is often (but not always) highly political because how we understand policy causality can shape what decision alternatives are considered as plausible or desirable and which are not. Consider the ongoing debate over the origins of COVID-19. That debate is highly politicized for reasons of its implications for geopolitics and research regulation. Explicit association of claims of policy causality with recommended policy actions holds potential (to some degree) to depoliticize our policy debates.
Making sense of policy causality in specific settings, such as in the aftermath of a disaster, is one area where policy research holds great potential to contribute to better decisions — those more likely to move us toward achieving our collective goals. However, understanding policy causality can be challenging, both as a technical endeavor and as a subject ripe for politicization.