Disaster Risk Management

The other car can be considered to be the hazard. There are some lunatics out there, and the frequency with which people are killed on the road in this way is well known, as is the impact when they hit other cars head on. You had prepared for this accident – but it was, of course, no ‘accident’ … it was a disaster waiting to happen – by choosing a car with airbags, and by having a first aid kit available. You had further mitigated the likelihood of being thrown through the windscreen by wearing a seat-belt. The authorities, well aware of such risks, had further mitigated the potential for avoidable death and suffering by imposing speed restrictions, putting up crash barriers and painting white lines on the road. Your passengers were not so lucky. They were exposed to exactly the same hazard, but lacked protection and so were more vulnerable. The authorities, knowing the evidence, had prepared and had an ambulance there in minutes to whisk the bodies away to a cold store at the morgue, all neatly tagged and wrapped in body bags. Such risks could, of course, be prevented by banning driving altogether!

But driving isn’t banned, even though it constitutes one of the largest threats to public health in developed countries. And it isn’t banned because each of us computes our own level of ‘acceptable’ risk. We make trade-off’s every single day between our perceptions of hazard, probability, and vulnerability to come up with what we consider to be an appropriately managed solution to the risk given the potential ‘reward’; in this case time saved by not having to take the bus, for example.

Do we really know enough, though, to allow us to make this calculation? Did the people of Christchurch, New Zealand, know on the 22nd February 2011 when an earthquake devastated their town that eight-storey buildings collapse in earthquakes more often than those that are shorter or taller? Would it have made a difference to their daily lives if they had? Is it more important in a flood to know how to swim, or know what to do if bitten by a snake? (The answer is very much the latter, which might surprise you, as comparatively few people drown in floods while many more die of snake bite). Is it important to know that over 700 people were pulled from the rubble of their homes after the earthquake in Pakistan in 2005 with spinal injuries who were able to move their limbs before they were ‘rescued’ but were paralysed by the time they reached hospital? Or that, without inoculation, tetanus has the potential to kill just as many people as falling debris?

If a massive earthquake were to occur in the middle of the largely uninhabited Gobi desert, would this constitute a disaster? Probably not, because nobody apart from a few hapless Yaks would be around to get hurt. A natural event does not necessarily lead to a disaster. So, we must infer that natural disasters are not ‘natural’ at all, but created by us. If, as with the total driving ban mentioned above, we were being entirely logical, we would also have to conclude that all disasters are avoidable since good ‘disaster risk management’ prevents avoidable loss of life, injury, damage, and livelihoods.

Furthermore, the fact that people suffer from the effects of naturally occurring or politically inspired events indicates a failure of development in its broadest sense – the sad consequence being a country’s failure to realise its true potential.

The 2010 flood in Pakistan is but the latest example, with meteorological warnings of record upland rainfall and Himalaya snow-melt unheeded, illegal de-forestation rampant, and flood mitigation through dredging and levée building paid for but undone.  This is another case of the “Samaritan’s Dilemma” where the humanitarian community is expected to step, literally, into the breach. But should humanitarian aid – and the neutrality and impartiality of the agencies dispensing it – compensate for the corrupted inaction of others going back decades?

Risk management is all about managing uncertainty, reducing vulnerability, and building resilience for communities at risk.

Against this sort of backdrop Clusters are asked to mainstream ‘disaster risk management’ into their response strategies in an attempt to minimise the risk of humanitarian action inadvertently having a negative impact on longer term development. Some might think this somewhat disingenuous, and an example of ‘risk transfer’ in its own right.

Mostly, these risks are programmatic and include such things as: the distortion of local markets with ‘free’ foreign inputs; substitution of state functions; compounding of ethnic, religious, and gender discrimination; and dependency creation. In addition, failing to link humanitarian objectives to developmental approaches (I hesitate to use the word “agenda’s” as that implies, by definition, politicisation) may present risks to the sustainability of assistance, and the building of capacity and resilience among disaster-affected communities.

Clusters increasingly get involved with other aspects of ‘disaster risk management’, too, particularly multi-hazard risk assessment. Haiti, as a country beset by earthquakes, hurricanes and floods is a good example of this. It was clear throughout the response phase that the recovery phase would bring with it its own series of challenges, not least disease outbreak (the cholera outbreak of August, some eight months after the earthquake was predicted and partially planned for), new earthquakes, landslides, mudflows, drought, and even tsunami (four people were killed by a mini tsunami on 12th January in Leogane).

As a result, a vision for risk management, including coordination of the response, was centred on a multi-hazard perspective which was meant to serve as a platform for ensuing risk assessments. This was aimed at understanding and communicating risk, assisting political and Cluster decision-making for land use planning, and reducing and transferring risk. These, after all, were the pillars for reinvigorating national risk management and development planning policies most of which were led by the ‘cartiers logements’ sub-Cluster within Shelter.

According to ODI, risk is perhaps best understood in terms of the concept of future harm; the probability of a harmful event or hazard occurring, and the likely severity of the impact if it does. Risk management refers to attempts to remove or at least reduce risks of future harm. It does this by

• Identifying potentially harmful events or hazards
• Assessing the probability and likely impact of these events
• Preparing appropriate prevention  and mitigation responses, and
• Real-time monitoring so that early warning can be given where possible

The measures through which it is possible to manage risk includes any combination of three strategies:

• Avoidance: Relocating people deemed to be living in areas of unacceptable risk (remember the Indonesian military forcible removing farming families from the slopes of an erupting Mount Merapi?); or ceasing activities that induce or exacerbate risk (such as girls’ education programmes that result in their fathers being shot)
• Reduction: Taking action to eliminate the probability of an event occurring, or mitigate the impact if it does. It can also involve sharing or transferring part of the risk through, for example, micro-insurance schemes or managing programmes remotely with national staff.
• Acceptance: Accepting residual risk, but only when based on utility and social cost benefit analysis rather than on biased perception and/or the idealism of the ‘humanitarian imperative’ [see section on ‘costs & benefits].

There are fundamental differences in approaches to managing risk between development and humanitarian organizations and individuals, even when they are engaging in the same context. The military engaged in so-called ‘stabilization’ operations find this a difficult concept to understand. Development actors commonly assess the risks of engaging in a particular context or programme, whereas the emphasis in the humanitarian world is on the risk, or human ‘opportunity’ cost, of not engaging.

It  is also commonly assumed that that there is a degree of creativity, innovation, and risk-taking behavior involved in humanitarian action – what the uninitiated refer to disparagingly as the, “cowboy” approach – compared with  the more cautious approach taken in the name of accountability by typical development actors.

Effective risk management is underpinned by rigorous assessment of contextual, programmatic, and institutional risk. This is the basis of ‘enterprise risk management’ models.

But humanitarian action takes place in a dynamic, complex, and complicated environment [see section on ‘complexity’] which means that time pressure, scarce human and/or financial resources, and the overwhelming need for immediate assistance make detailed situational risk analysis either impossible or not a priority. People also perceive risk differently which makes all risk relative. Risk is also relative in the sense that it is a function of uncertainty, the fact that probabilities for different hazard events are either changing or the underlying ‘risk factors’ are not well understood. ‘Disaster epidemiology’ makes use of whatever evidence is available to reduce this uncertainty so that we understand better what turned a phenomenal event into a disaster so that we can reduce risk next time.

The four types of risk analysis that should be carried out if possible, however, are:

• Contextual Risk Analysis: to understand relative risk, including those that are outside the agencies’ control. For example, annualized per capita aid expenditure for earthquake victims in Haiti is $1,492; in Darfur, $230; and $32 for the Pakistan floods
• Programme Risk Analysis: clarifies and prioritizes objectives in terms of identified hazards, risk factors, absolute need, gaps, and capacities. Wherever possible, analysis of hazard should be multi-hazard as, for example, earthquakes can set off land-slides which in turn cause floods, and cholera epidemics can result from inadequate sanitation.
• Vulnerability Risk Analysis: Vulnerability is a function of three interactive components: Exposure (e.g urban migration, land use practices, population growth); Sensitivity (e.g poverty, age, disease profile); Resilience (ability to adapt behaviour in order to cope better)
• Institutional Risk Analysis:  This includes reputational risk to an organization. Lessons learned from the Haiti and Pakistan mega-disasters in 2010 demonstrated once and for all that failure by some of the larger Global Cluster Lead Agencies to ‘mainstream’ or take the Cluster Approach seriously had serious negative consequences for institutional fund raising and, by extension, the dispossessed they were there to service in the first place.

On the other hand, the Cluster Approach has tried valiantly to address the management of programmatic risk by producing common inter-agency (IASC) guidelines; investing in comprehensive, risk-oriented needs assessment and monitoring systems; providing accessible stockpiles; and, most important of all, strengthening Cluster and Inter-Cluster coordination so that the risks from inappropriate or duplicated aid delivery, as well as gaps in aid delivery, are minimised.

The last point that needs making concerns the confounding role played by ‘perception’ on probabilistic modeling. This story, borrowed from the World Bank, says it better than I can:

“During the next ‘solar maximum’ there is a very strong possibility that we will experience one or more solar flares that have enough energy to disable or destroy our entire Global Positioning System (GPS) satellite network [ – and all of us with it, if you believe Nicolas Cage in the film ‘Knowing’]. Because that network is now being used in a way that was never intended, namely to synchronize power generation throughout the US, the loss of that network might well bring down the entire electrical grid of North America.