A mailing from the Royal United Services Institute invites me to a conference in April:
The Critical National Infrastructure (CNI) is both the backbone and the lifeblood of the country. It comprises the assets, services and systems that support the economic, political and social life of the UK. Any disruption, damage or destruction to all or part of the CNI could result in grave consequences for the functioning of government, the economy and society. Clearly the CNI is vital to the country’s well-being but the planning and implementation of its security is a Byzantine process; the CNI is a complex and uneven environment with ownership and responsibility spread across the public and private sector.
The threats it confronts are myriad including terrorist attack, industrial accidents and natural disasters. As demonstrated during the July 7 bombings, the Buncefield Disaster, and the foot and mouth outbreak, the CNI is a labyrinthine web of interdependent vulnerabilities that requires a coordinated and coherent response across its entirety to ensure its effective security and resilience in the face of such threats.
Dangerous rubbish. This is an epitome of the statist miscomprehension of complex systems, of economies and ecologies. ‘It is messy; we must coordinate it,’ they say. There are vital things that can be identified in advance as such, and other things not necessary to the ‘backbone of the country’, they think.
But the connections in a natural web are flexible, or they don’t get established in the first place. “Interdependent vulnerabilities” are what make systems adapt, the source of resilience. In unmanaged, open, systems everything is important and everything is unimportant: all things contribute their part to everything else (and you can’t directly measure their contribution), but competition ensures they are all redundant and replaceable.
The response to 7 July was a demonstration of improvisation by thousands of separate actors – millions if you count all those who took simple decisions to get out and walk, rather than passively waiting to be evacuated by the authorities, which would have been the orderly, planned, way to do it. London was functioning again in a day, despite, not because of, the “strategic interventions” that restricted the recovery of traffic flow, and filled the streets with police.
Livestock farming in Britain almost didn’t survive the Deprtment for Rural Affairs’ “coordinated” response to the last “foot and mouth” outbreak. Fortunately at the time DEFRA lacked the powers to coordinate more farmers out of business. The department didn’t see it like that: Its plans were frustrated, and that’s why things were as bad as they were. The ‘defect’ has been eliminated by the Animal Health Act 2002 and the Civil Contingencies Act 2004.
Nobody in government had to tell Tesco’s dealers to buy up more petroleum in Rotterdam when the Buncefield depot caught fire. The state way is a ‘strategic reserve’ of petrol under armed guard somewhere, distributed eventually by rationing according to who is important enough to get it, after declaration of a suitable emergency. As it was, loss of 20% of the country’s stocks overnight caused scarcely a single car journey to be cancelled – apart from those of the people no longer commuting to the flattened industrial estate.
Those ex-commuters would not be comforted by the thought that distributing tiles or soft drinks is not “critical” and not to be guarded by the state. What they do matters to them and their customers. When I want petrol, petrol matters; when I want tiles, they matter. We are all equally made poorer by the unavailablilty of either, because we can’t predict what we will want. Nor can the state.
How dare the planners decide for me what it is I want, as they do implicitly when they define some workers, some structures, as “key”? Well there’s a confirmation bias at work. What the state can best monitor is important (invisible, uncontrollable processes couldn’t be); so those who work for it are. Chaos is bad. State plans are designed to control chaos; therefore they do, and any unfortunate or unforseen consequences are just the remnants of chaos uncontrolled. Bad things are not in the plan, so not of the plan. They are part of the failure to squeeze out doubt, never caused or exacerbated by wrong or unnecessary decisions by the authorities.
The misunderstanding at the heart of planning is a fundamentalist belief that order and simplicity are public goods. They aren’t. It may be good to have them in your own life – if you want them. It is probably necessary to have them in managing a task, running a business, playing a game; to make any well-defined single goal attainable. Clarity in shared procedural rules is highly desirable. But if we want to live in a world where the goals and threat aren’t well defined, where we have a choice, and where how we live is not vulnerable to simple shocks from unexpected angles, then universal order and simplicity are bad. Conflict and competition, difference and redundancy are good. The more disorder, uneveness, and complexity our society has, the richer our lives, and the better equipped we are collectively to meet disaster by routing around damage.
Each entity should be responsible for its own contingency.
Buncefield – aptly named – is an example of when an entity is NOT – i.e. they had little to zero resources for dealing with a fire at their plant, but sprawled, boneless, waiting for the local Fire Brigade to pitch up.
If the RUSI want to do good, it is just to raise awareness in entities of their own personal responsibility for any response. Who else would know better about each facility than the operatators themselves? If they are unclear, they should contact others in similar circumstances to share wisdom and form alliances as and when both see fit. The point is they are responsible, but it does not mean that each has to provide every aspect of implementation in isolation.
The Government can and I believe should play a part in raising awareness. When they get actively involved in a nannying, “Oh! give it here!”, dinnerlady-cutting-up-food sort of way then we see the abdication that occured at Buncefield.
An excellent posting all round from Guy. I particularly liked:
A couple of supportive comments.
(i) Disaster planning is largely about avoiding a single point of failure actually being critical. If overdone, central control of the plan, itself, can introduce a single point of failure.
(ii) The Internet is a good example of a robust system. There is no central plan, just a fair bit of redundancy.
In fact planned redundancy is the key to robustness. Its lack is one of the reasons why the NHS is falling apart.
Anyone can sqeeze more short-term efficiency out of a system with protective redundancy. Only real managers, steeped in their field of operation, know how to balance that against longer-term needs.
Best regards
Great post! Have you have read “An Army of Davids” by Glenn Reynolds or read a precis of his book on the web? It makes broadly the same point that human societies, precisely because they are “messy”, are often far more flexible in the face of problems than centrally-planned ones.
And of course Hayek et al made these points decades ago. Come to that, Edmund Burke did so 200 years+, if one thinks about parts of his critique of the French physiocrats.
I take it you weren’t invited to speak at this conference Guy? Now that would be something to see.
John Reid appears to be the main speaker. I’m fairly sure RUSI does not want a matter-antimatter annihilation in the middle of Whitehall (however good that might be as a general idea).
Good post! RUSI are one of those organisations that make up potential crises so that they can hold conferences about them (capitalist exploitation of the market for PANIC!!!) but it probably wouldn’t be a bad idea if you were to go and tell them. You might make some people more thoughtful, even if you don’t make much impact on the views of those selling doom.
The connections in a natural web start off flexible, but there is such a thing as over-adaptation. When the environment is stable for a very long time, systems specialise to take maximum advantage from subtle economic/ecological niches, and lose their generalist adaptability. It happens most with island ecologies, where birds lose the ability to fly and the sense to run in the face of a complete absence of predators, and then fare very badly when the people and their pets arrive.
The thing that keeps a web adaptable is the constant noise of knocks and changing circumstances. You lose out on the advantages that specialisation gives you, which are considerable. But if you want the web to be strong against outside invaders and other disasters, the best thing to do is to get in its way. If someone does something clever, block it, so they’ll have to find another way round. Make stupid regulations, and then change them, and then change them again, so people are experienced at setting up new ways of doing things. Shut main roads for roadworks at rush hour, so everyone who commutes knows every side-road and back road around every part of their journey. Screw every single system thoroughly until everybody spends half their day practicing the building of new networks and routines from scratch to bypass the snafu you just made.
That way, when the disaster happens, nobody will even notice it against the background roar of big government. Resilience against disaster only evolves in the face of disaster, so an intelligent planner would spend his time dreaming up ways to create some.
… I wonder if they’ve ever thought of that?
I’ve just been reading Dr R V Jones’s ‘Most Secret War’, about his experiences in scientific intelligence during WW2.
Repeatedly he stresses the need for a small-scale flexible organisation to deal with the various stands of intelligence coming in, and a recurring theme is his resistance to having his own organisation “co-ordinated” and subsumed into the War Office monolith
I think it’s true to state that optimisation of any system leads to a fragile design that is likely (when it fails) to fail catastrophically.
In an emergency, hang every civil servant in sight. This should reduce the system’s load to an acceptable level, and will have no affect whatever on its performance.
Except, perhaps, for a transient shortage of rope.
In an outbreak of bird flu, we could confine this to merely hanging every vaccinated civil servant. The populace would then be protected much more efficiently.
Tch. Now any civil servant will quite properly point out that this is an appallingly inefficient waste of rope, and didn’t you know you can re-use the stuff. You need to get this thing organised and coordinated if you want to get anywhere…
Now, have you thought about what you’re going to tie the rope to, where you’re going to put the bodies (there are H&S issues to consider, you know), how you’ll arrange transport to the gallows for several tens of thousands of public servants, and how you’ll process the travel expense claims through the estates of the civil servants hanged in order to ensure all the books balance…
Now, let’s set some targets for number of civil servants hanged per year, a couple of forms to fill out crossed checked against payroll records just to make sure we’re getting them all, and put form 12948… 321… (b)… part C (or was it part B? No, C I think) in to requisition the appropriate amount of rope, gallows, cattle cars, additional “Arbeit Macht Frei” signs, a form 91325 requisition for extra copies of form 386 – travel expenses (deceased) – and all the H&S clearances into procurement branch, we’ll have the whole thing done in a jiffy…
Oh yeah, and form 623 to report yet another instance of affect/effect confusion to the dept of statistics (Education (linguistics)) branch… 😉
[rolls eyes heavenwards] Transient shortage of rope…
Pa Annoyed wrote:
Two thoughts:
(i) If you keep your tongue in your cheek too long, you risk biting it;
(ii) Said policy perhaps protects against the internal problems, but not the external: in particular the long-term economic threat from Asian tigers. [Also, noting today’s news, how big is the Latvian navy?]
Best regards
Aren’t Asian tigers on the endangered species list?
If you don’t laugh, you’ll cry. And to take the point seriously, there is only any point in having a navy if you plan to use one, and there is only any way to fund it if there is public support for such use.
For whatever reason, there is now no public support for the actual use of our military, and greatly reduced public support for the military itself. Having tried to motivate more support with a war for which we are woefully and visibly ill-equipped, the utter failure to materialise of any significant demands for a greatly expanded military capability leaves them with no other option.
Appeasement is the latest fashion. Guns are un-hip, bombs are square, and warships are just sooo 1980s!
But this all emphasises exactly the point I was making. Thinking we need a large military is the thinking of centralised long-term planners, coming up with the sort of scenarios RUSI specialises in. Adaptable markets say we don’t use them, so we don’t need them. We live on an isolated historic island of peace, and we are about to lose the ability to fly. (Float. Whatever. You know what I mean.) Natural selection says the only way to keep it requires facing dangerous predators. We need those tigers.
People can get on without planners a lot better than planners can without people.