Draft Review of Cerulo
My draft review of Karen Cerulo
is over the fold. Comments much appreciated
In late August 2005, I was visiting the United States, and experienced, along with hundreds of millions of Americans, a growing sense of disbelief and outrage at the governmental response to the destruction of New Orleans by Hurricane Katrina. This was a disaster that was both predictable and regularly predicted, yet days after it struck the combined efforts of federal, state and local governments were insufficient to evacuate thousands left homeless and in desperate straits. Similarly, as has now become evident, the US government had received regular warnings of the likelihood of an attack like that of September 11, 2001, but did little or nothing to prevent such an attack.
The lack of coherent planning evident with Katrina raised all sorts of issues of race, class, partisan politics and ideology, but the most striking impression was one of bafflement. It seemed that, despite ample warning of the potential disaster, those responsible never saw it coming. Much the same could be said of 9/11.
Karen Cerulo in a book aptly titled Never Saw It Coming:Cultural Challenges to Envisioning the Worst argues that the failure to prepare for disaster is not a matter of individual incompetence or fecklessness. Rather she argues it reflects a bias towards optimism that is deeply embedded in American culture.
Ceruloâ€™s argument centres on the concept of â€˜positive asymmetryâ€™, in which best case outcomes are seen as more representative or normal than worst case outcomes. Positive asymmetry is closely tied to core American cultural values, most obviously â€˜positive thinkingâ€™. The injunctions to â€˜be all that you can beâ€™ and â€˜expect the bestâ€™ lead inevitably to collective miscalculations.
Cerulo argues that this positive asymmetry is evident both in individual decisions and in the decisions of small and large groups. In all cases, she suggests, the asymmetry is produced by adherence to cultural norms. On the other hand, Cerulo argues that deviant groups are treated with the opposite asymmetry, with an undue focus on negatives.
Ceruloâ€™s book may be located as part of an emerging American literature in support of pessimism. Clearly, this literature is driven in part by a reaction against the disastrous outcomes of recent American over-optimism. Whereas Joshua Dienstagâ€™s Pessimism:Philosophy, Ethic, Spirit presented a philosophical case for pessimism, Ceruloâ€™s argument is sociological. Compared to Dienstag, Ceruloâ€™s argument is more specific to the United States, but both are ultimately arguing against optimistic and progressivist traditions derived from the Enlightenment.
The most important distinction between Dienstag and Cerulo is that Ceruloâ€™s sociological perspective leaves open the possibility of structural changes in social institutions that would eliminate or reduce positive asymmetry. Cerulo identifies some professional cultures which she says are less prone to positive asymmetry (notably public health and information technology professionals) and notes some apparent success stories in preparing for the worst, such as SARS and Y2K.
Demonstrating that no-one can truly escape their cultural background, Cerulo ends on a characteristically American note of hope and exhortation, saying that â€˜we can do betterâ€™ and â€˜we would truly be remiss if we simply failed to tryâ€™.
In the abstract Ceruloâ€™s argument seems convincing, and there is plenty of evidence to support it. But there is lots of room for disagreement over details..
Cerulo criticizes the authors of baby-care manuals for devoting insufficient space to developmental disorders and other things that might go wrong and for following up such information with reassurances that serious problems are unlikely. She doesnâ€™t, however, suggest any useful improvements.
Indeed, Cerulo is particularly critical of the most widely-used response to concerns about developmental abnormalities, amniocentesis, pointing out that the risks of this procedure outweigh the potential benefits in many cases, and criticising writers on the topic for downplaying these risks. One gets the feeling that the writers of these books canâ€™t win. If they play down the risks, they are accused of not preparing for the worst case, but if they play them up they run the risk that more expectant parents will seek to allay their concerns through medical procedures that carry their own risks.
More strikingly, one of Ceruloâ€™s examples of unfair treatment of deviant groups is the Heavenâ€™s Gate cult which, as she states believed that they would be removed from the Earth by a spaceship following the comet Hale-Bopp, their true homeâ€™. As she says, most reporting of the group treated it as the epitome of the lunatic fringe. But readers willing to go to the endnotes will learn that this treatment might be explained in part by the fact that in order to hasten their arrival in heaven, all thirty-nine members of the group engaged in a mass suicide. (While Iâ€™m on this subject, is there any excuse for persevering with endnotes? If the material is of too little interest to be included in the main text or in footnotes, couldnâ€™t it be placed in a supporting website or omitted altogether).
Another surprising claim is that â€œpositive asymmetryâ€? is demonstrated by the fact that, in theology and art, Heaven is given a detailed and appealing description, while hell is described only in vague and non-specific terms.On the contrary one of the standard criticisms of religious art is that Hell and the Devil are made much more interesting than Heaven and Hell, notable examples including Dante and Milton.
On the other side of the coin, Cerulo cites as an example of successful preparedness the massive Y2K remediation effort undertaken in the United States, and devotes quite a few pages to a largely uncritical presentation of the â€˜officialâ€™ view that the remediation effort prevented an otherwise inevitable disaster.
The official view doesnâ€™t stand up well to a close examination of the evidence, which suggests that the severity of the problem was grossly overstated. Many countries undertook little or no Y2K preparation and came out fine. Russia and Italy are notable examples – the US State Department issued a travel advisory for Italy as did UK authorities. Australia actually evacuated its embassy in Moscow leaving a skeleton staff to wait out the cataclysm. This isn’t a matter of being wise after the event. Once the 2000 fiscal year began with no serious incidents it was obvious, and was pointed out for those willing to listen, that for anyone except nuclear reactor managers and the like, ‘fix on failure’ was the optimal response (Y2K bug may never bite, Australian Financial Reviewm 2 September 1999
Cerulo quotes Y2K alarmist and entrepreneur Peter de Jager as an authority. De Jager was a leading proponent of the ultimate worst-case scenario, that of widespread failure of â€˜embedded systemsâ€™, hardwired chips built into cars, aeroplane navigation systems, lift systems and the like. Since, as rapidly became apparent, these systems could not be reprogrammed, the only options were to scrap them and, in most cases, the systems of which they were part. Fortunately, precisely because they were hardwired these systems neither had nor needed the capacity to maintain an internal date – it would be lost the first time such a system lost power. Once these points became apparent, mainstream Y2K boosters quietly forgot about embedded systems, but they remained a core article of faith the â€˜Y2K as Armageddonâ€™ movement until the uneventful dawning of the fateful day.
Itâ€™s hard to know what to make of these problems with the details. Some disagreements is to be expected in any detailed argument. Or perhaps the problem is that handling low-probability catastrophic risk is something we are not very good at, sometimes preparing for non-existent risks and at other times failing to foresee obvious possibilities.
There is, however, one significant gap in Ceruloâ€™s analysis, which could usefully be filled. Although the discussion of envisaging the worst is extensive, and generally sound, Cerulo does not put forward, or even refer to, any explicit framework for making decisions in the presence of low-probability extreme risks. The voluminous literature on risk analysis is simply ignored.
It might be useful to put forward a highly simplified summary of the way in which a decision analyst would approach the kinds of worst-case scenarios described by Cerulo.
First of course it is necessary to identify the worst-case scenarios (An exercise in risk management: on what to be scared of and what to do about it, AFR, 10 September 2004). The next, and essential, step is triage.
Some putative risks turn out, on more careful examination, to be so improbable as not to justify further attention. For example, while shark attacks make dramatic news, the risk faced by any individual beachgoer is tiny (there are between 30 and 100 attacks per year worldwide), and far outweighed by more prosaic dangers such as drowning and skin cancer.
On the other hand, there are risks that cannot be reduced by anything we do. For example, the lifetime risk to each of us of dying from a massive asteroid impact of the kind that wiped out the dinosaurs has been estimated at one in 50 000, comparable to the risk of dying from snakebite or lighting strike. However, whereas we can take sensible (and low-cost) measures to reduce the risk of snakebites or lightning strikes, there is, at present, nothing we can do about asteroids. So, it makes sense to ignore this risk in making day-to-day decisions.
The remaining, decision-relevant, risks are those large enough to matter, and manageable enough to permit a response. Within this range, economic tools like benefit-cost analysis come into play. It rarely makes sense to attempt to completely eliminate risk. Rather the optimal response is to reduce decision-relevant risks to the point where they are either too small to worry about or where there are no remaining options that achieve a reduction in risk that justifies their costs. Of course, there is lots of room for disagreement about costs and benefits as well as about the specifics of particular risks, so the problem is unlikely to be solved any time soon.