Friday, May 29, 2009

Lessons in public administration from the Victorian fires

I haven't commented on the on-going inquiry into the Victorian bush fires because of lack of knowledge of the detail. However, there was one story from yesterday that caught my ear because it seemed to me to be a symptom of a growing administrative problem that I have talked about before.

The facts (and here) appear to be this.

Fire hit Kinglake around 6.25 in the evening. Just after 1.30 that afternoon, a mapper at the Kangaroo Ground incident control centre predicted that fire might strike as far as Kinglake and St Andrews. By 3 an urgent threat message was ready to go from the Centre. Kangaroo finally issued a warning at 5.20, then a warning appeared on the Country Authority Web Site at 5.55. By then it was far too late.

But why the delay? I quote:

Today the senior counsel assisting the Royal Commission, Jack Rush, questioned CFA operations manager, Jason Lawrence, who was in charge at Kangaroo Ground on the day.

JACK RUSH: Who says that you can't issue warning messages if you haven't got control of the fire?

JASON LAWRENCE: Well the warning messages are to be issued by and signed off by the incident controller and as I wasn't allocated that role or unable to perform in that role at the time, then I was not able to sign off on those information releases.

SIMON LAUDER: Mr Lawrence told the Royal Commission he tried several times over the afternoon to contact those who did have the authority to issue warnings but he couldn't get through.

JACK RUSH: And Mr Lawrence, that wasn't released because you couldn't contact the Kilmore ICC.

JASON LAWRENCE: Well it wasn't released because once again I wasn't the incident controller and not responsible for the release of that information but also in part because I couldn't contact the ICC up there to confirm that and hold those discussions.

Kangaroo finally issued a warning without authority after learning that communications at the Kilmore Incident Communications Centre were down.

Before we draw any conclusions about Mr Lawrence's failure to act independently and release the information, we should remember two things.

The first is the chaos of that afternoon. The second is the way the rules bound him, reducing the capacity for initiative. They did try to contact Kilmore ICC a number of times, but simply could not get through.

One point that I have made many times in writing on today's systems of public administration in Australia is the way they have become more mechanistic, more controlled, more rule bound, more rigid. They have been reborn as the human equivalent of a computer system, parts that are meant to work together in defined ways following defined chains of communications and operating under specified decision rules.

This system can be quite effective when working with the known, although I have pointed to specific problems here. It struggles when it has to deal with the new, or with fast moving events that require action outside the rules.

You can clearly see all this quite clearly when Mr Lawrence says:

Well the warning messages are to be issued by and signed off by the incident controller and as I wasn't allocated that role or unable to perform in that role at the time, then I was not able to sign off on those information releases.

Note the words incident controller. It is obviously sensible in managing responses to an event like the fires to have a degree of central control. However, the very words "incident controller" carry with them a tone that carries through in the idea that warning messages must be issued by and signed off by the incident controller. Mr Lawrence had a place in the chain of command and, in terms of the rules applying to him, he was not allocated the role, could not perform the role.

It seems clear from the evidence presented so far that, to a degree at least, the Victorian attempts to control and manage such a vast outbreak of fire collapsed under systemic failure. In particular, the communications systems on which central command and control depends failed. Once communications go, then the system fails.

In military terms, the phrase fog of war was coined to describe the way in which battle field ambiguities and difficulties affect action. One central control fails or becomes difficult, then you need to rely on the initiative and actions of individual units. The system itself can then become an impediment.

In considering Mr Lawrence's actions, you have to ask whether you would have acted differently. I don't know that I can say in all honesty that I would have had I been working in that system at that place at that time. The chaos and pressure must have been quite dreadful.

One of the things that worries me about likely responses to the Victorian fires is that attention will focus on ways of making the system work more effectively, of adding further layers to the system to overcome the problems that arose this time. Maybe the system itself should be abolished, replaced with one where planning starts from the local and builds up, with the central role redefined as facilitating and coordinating, not controlling.

A brief postscript

I was chatting with colleagues this morning about this one. Just a few brief comments on points made.

Once you take way or in some ways constrain individual freedom and responsibility, you cannot then expect people to re-assume it automatically when systems fail.

The role of insurance, fault and litigation. Should we adopt the New Zealand no-fault system?

Then the conversation switched to rules and unforeseen events in a more general sense, the way in which policies and official structures created a world of their own whose linkage with actual life could become very tenuous indeed. There i another post lurking around here. I will leave it there.

No comments: