Category: Systems
“Charbonneau Loops” and government IT contracting
‘Charbonneau loop’ turns out to be one of those terms which we didn’t know we needed until it was called into existence, but draws attention to something all too easily overlooked. It describes a form of moral hazard, which is simple, obvious, and largely invisible:
Charbonneau Loops ultimately happen when the “pool” of companies (receiving public sector contracts for a given type of work) is small enough that the same companies are sometimes overseeing, and sometimes overseen, by their peers in that same pool. Even if they never actually coordinate with each other – even if they don’t have any conversations whatsoever – they’re all incentivized to be a little bit less critical of each other as a result.
The concept takes its name from an investigation into construction corruption in Quebec, but it can clearly apply to sectors other than construction and far beyond Quebec. It can also be extended beyond the simple two role form of the loop. The story of Grenfell Tower is a multi-player version where a complicated set of public and private sector organisations carefully positioned themselves not to identify risks and not to be resposible for resolving them.
The question, of course, is what can be done to break the loop and restore – or perhaps create – conditions in which the institutional incentives act differently, supporting effective challenge, rather than muddled complacency. The simple answer, set out in the post, is to strengthen in-house capacity and to increase the pool of suppliers. But as so often, the harder question is how to get there from here, and how to avoid letting progress be undermined by regulatory revolving doors which create a form of personal Charobonneau loops embedded in the more institutional ones.
So long and thanks for all the bits
This is a valedictory post from the departing technical director of the National Cyber Security Centre, which leads off with a quantum state superposition joke. That might all suggest something of interest only to rather a niche audience. It is true that some of the points made assume a level of familiarity with cyber jargon which not all of us possess, but that turns out not to matter nearly as much as it first appears. The post is in fact a set of ten lessons learned, all of them relevant much more widely across government (and well beyond) than just the immediate world of cyber security. Taken together, they read as a kind of manifesto for systems thinking.
As just one example, point 7 is that ‘incentives really matter.’ That’s something which governments haven’t tended to be very good at, perhaps in part because they don’t think they need to be – governments can, after all, make things mandatory, which is not an approach available to others. But in the end, actors in a system will behave as they perceive it to be in their interests to behave, and it is foolish to assume otherwise. As between government and technologyy infrastructure,
we implicitly expect these companies to manage our national security risk by proxy, often without even telling them. Even in the best case, their commercial risk model is not the same as a national security risk model, and their commercial incentives are definitely not aligned with managing long-term national security. In the likely case, it’s worse.
That that is true much more generally is both pretty self-evident and widely overlooked, and some pretty terrible things have happened as a result.
Carousel
This is another perspective on the (unconfirmed) news of another attempt at civil service reform, unpicking why simplistic approaches are doomed to failure. As is so often the case, treating an aim defined in one way (“the civil service will be 20% smaller”) as if it were the same as a quite different aim (“the civil service should become more efficient”) leads to hopeless confusion. Or as Mark puts it:
Efficiency may lead to the need for fewer staff but fewer staff does not lead to greater efficiency.
This post also underlines in a different way why it is so important to look at all this from a systems perspective. Government can – and does – operate at different levels and different configurations, and there is no reason to think that, in a world where permanent secretaries worry about fixing school boilers, the current balance is anywhere close to optimal. Focusing on a single layer can, at the very best, result in optimising that layer; it cannot result in optimising the system.
Civil Service Reform – Lord Maude Tries Again
Martin Stanley – UK Civil Servant
‘Civil service reform’ is an unintentionally revealing phrase. Its use is a strong indicator of somebody who hasn’t thought through what problem they might be trying to solve, still less what actions might lead to solving it. That’s not because civil service reform is not necessary or not desirable – on the contrary, it is very necessary and very desirable. It is because the civil service (itself a huge collective noun, concealing variety at least as much as describing a singular entity) is part of a wider system. Honest reformers recognise the need to address that wider system; rhetorical reformers do not always feel the need to do so.
Prompted by press coverage suggesting that Francis Maude might be about to be invited to have a third attempt at civil service reform – and with the primary success criterion clearly being the extent to which the civil service ends up smaller as a result – Martin Stanley patiently explains why Maude’s first two attempts failed and why any third attempt is unlikely to do any better. He lists nine problems consistently identified in past reviews of the civil service, all of them depressingly recognisable. But what is perhaps most striking about the list is how much of it is rooted in what ministers and Parliament do (or don’t do) and how little of it is limited to what the civil service does in isolation from that wider system.
Again, that’s not an argument that all is well in the civil service or that nothing there needs to change. Almost exactly ten years ago, I wrote a blog post on this issue, prompted by the civil service reform plan published in Maude’s name. It is not reassuring that the last paragraphs of that post seem just as apposite today:
The civil service is big and complicated and there are important ways in which it could change for the better. But big and complicated as it is, it is also just a component of the wider system of government. The more radical the ambition for the civil service, the bigger the implications for that wider system will be.
If you want to change the system, you have to be ready to change the system.
The Wetware Crisis: the Thermocline of Truth
This is an old post, but a good one. I had occasion to look it out again this week, so posting it here both as a service to others and to make it easier to find the inevitable next time I want to show it to somebody.
How is it that projects seem to be steadily on track for successful delivery until just before the target date for implementation – at which point they suddenly aren’t? It seems implausible that the phenomenon should be helpfully explained by reference to temperature gradients in oceans, but that is exactly what this post does. For all sorts of reasons – some covered in the post – reporting systems give false assurance to senior decision makers for far too long, until the tipping point is reached at which the actual state of readiness can no longer be concealed, when suddenly the project dashboard goes from benign to catastrophic in an instant. Crossrail provided a particularly spectacular and very public example when it failed to open in 2018 having been presented as being on track to do so only a few months earlier – despite what we now know to be a need for more than three years further work before the line could open.
But this is not a phenomenon limited to big infrastructure projects. I can readily think of examples at more or less every scale, and I suspect anybody with any familiarity with project delivery will have their own. From that experience, I see another critical factor encouraging the development of thermoclines of truth: the setting of delivery deadlines. The further ahead the deadline is set (and the more arbitrary it necessarily is as a result) and the more visible that deadline is, the more certain it is not just that the deadline will not be met, but that the fact that it will not be met will only be recognised at a very late stage. There are some pretty obvious reasons why this adds to the risks Webster identifies in his post – the stronger the commitment, the more visible the failure, and the greater the reluctance to face up to the problem, still more so to be the messenger of bad news.
Delivery happenes when the work is complete. Completion of the work does not happen because delivery is due.
Beyond the Web
Metaphors have power, they frame how we think about things, sometimes rather insidiously. We talk blithely about ‘the web’ as if it were a thing and as if that thing were well described as a web. But just as the cloud is somebody else’s computer, so the web is everybody else’s computers, and the connections between them are much less a web than a tangle or, as Searls puts it, a haystack.
Searls is one of the original cluetrainers, and so has been thinking about this for a long time. His metaphor is, of course, no more an accurate description of anything tangible than the conventional one we are all familiar with. But changing the metaphor changes some of the questions in some challenging and important ways, and eleven rather good ones are listed in his post.
The web we all currently experience is a set of points – points which may be linked together, but still points. It is driven by a client-server architecture which takes our data and scatters it around in ways were cannot control, but which is structurally incapable of delivering integrated experiences in return. That’s so normal that we barely recognise it – but it’s not the only possible way for systems to be connected and for data to flow.
So yes, the Web is wonderful, but not boundlessly so. It has limits. Thanks to the client-server architecture that prevails there, full personal agency is not a grace of life on the Web. For the thirty-plus years of the Web’s existence, and for its foreseeable future, we will never have more agency than its servers allow clients and users.
It’s time to think and build outside the haystack.
The (il)logic of legibility: why governments should stop simplifying complex systems
Thea Snow – Centre for Public Impact
The ability to abstract, to stand back, to see the large pattern obscured by the detail is an important and powerful one for any strategist. It is also a very dangerous one.
It is too easy to impose a pattern or to assert structure which is not really there. If we abstract from messy reality, we may be able to get rid of the messiness, but in doing so we are all too likely to lose contact with the reality as well.
To acknowledge that, to recognise complexity and accept the uncertainty which it brings, is not a sign of weakness or intellectual inadequacy. On the contrary, strength and resilience – in reality and in the understanding of reality – come from accepting and embracing messiness and the strengths it brings.
This post celebrates complexity and the constraints on knowledge it imposes, and a world in which ‘I don’t know’ can be the most powerful thing we can say.
Joined-Up Thinking Requires Joined-Up Practice
It’s not hard to recognise that the world and its problems and opportunities are a complex system and that linear thinking and mechanical metaphors are not good ways of understanding and responding to that complexity. It’s also not hard to ignore that recognition and carry on as if connections were simple and systems comprehensible (though as the previous post argued, storytelling is a powerful tool to help us through that).
This post powerfully argues that joined-up thinking can never be enough, unless it leads to and is informed by joined-up doing:
If we want to change whole systems we’ve got to think and work as whole systems. Nobody can think non-linearly. None of us is that clever. The only way to think in a systemic way is together. Joined-up-thinking requires joined-up-practice. This is the meta-shift of our time, one that requires a new mind set and skill set: learning to think like a system by working as a system.
That of course poses an enormous challenge to the people, structures and processes of public policy making. This is not how problems are framed, still less is it how they are normally addressed. Changing the system of system changing without having already done it is not going to be easy.
The Inspector’s Dilemma
Public bodies do many things, not all of them necessarily at the front of mind when we think about what governments do and how those things might be done better. One of those things is, broadly, inspection – checking to see that requirement which should be met are being met (and sometimes to see whether failures reflect inadequate requirements or poor compliance). The existence of regulation and inspection raise an important question about the attribution of responsibility: does the very existence of a regulatory system shift responsibility inappropriately, does it in effect create a form of moral hazard? And if the answer turns out in practice to be to fudge the issue, the consequences may turn out to be very bad.
We need new operating systems. Whose job is that?
This is not a post about operating systems, in the obvious sense. It might be better described as a post reflecting on Marx’s early view that the base determines the superstructure, so that to achieve deep change your intervention needs to bite deep into the system. Many of our component systems are deeply rooted in the circumstances of their creation and evolution, they are solutions to the problems of their time which may not be our problems. But they are neither easy to change nor perhaps even easy to recognise for what they are. This post describes a clear problem, a future post offers the promise of an approach to answering it.
‘Government as a system’ for COVID-19
This post takes the government as a system approach which the Policy Lab has been developing and applies it to the policy challenges created by COVID-19, less in terms of the immediate response, more in terms of emphasising three areas where modern policy approaches are likely to be critical.
The first is thinking about the future in conditions of particular uncertainty. Doing that creatively, radically, realistically and usefully is hard enough at the best of times, and these are not they. The more self-consciously and the more collaboratively that is done, the better the chance that the results will survive contact with developing reality.
The second is data, critically recognising that it is not just a matter of collecting updated answers to existing questions, but of identifying new questions and the data needed to answer them. The Policy Lab analytics ladder provides a really useful framework for thinking not just about how approaches relate to high level questions, but how their relative emphasis will change as we go through the crisis.
The third is relating all that and more back to the whole system and providing leadership and direction – a reminder that policy is an approach and an activity for those who want to change the world, not for those who wish merely to observe it.
Slowdown Papers
Most writing on strategy is brisk, terse, focused. It tends to the abstract and the impersonal. The author is a creature of intellect but not of imagination – and sometimes is an institution or a group rather than being an individual at all.
So this collection of essays is striking in form as well as content. They are highly personal, they are expansive, at times even meandering. At one level, they are a series of personal reflections, starting with the Australian bush fires, moving through the coronavirus epidemic, to the greater challenge of climate change and on to the social and economic state of the world – and the language we use to describe all that, which itself powerfully constrains how we think about it. But this is also – and above all – systems thinking on a grand scale.
The existence of the coronavirus and its direct impact on human health are matters of biology. But pretty much everything else about the virus and its consequences, both immediate and for the longer term, are social and economic – and so intensely political – issues. It is always true that we have choices about how we think about those kinds of issues, but it is a choice rarely exercised in the practical world of public policy making.
Government agencies in almost all countries tend to use only one or two models with which to formulate policy. Further, they rarely gather evidence and analyses on different competing assumptions, or contradictory models, and then tend to build capabilities around the existing underlying models, rather than cultivate new approaches, potentially exposing previous judgements. This would tend to suppress nuance, but also inhibit the exploration of new trajectories.
In that, of course, there will continue to be much uncertainty, bringing to mind the line that:
Democracy is the form of society devised and maintained by those who know they don’t know everything.
But to know we don’t know everything is to know something very important, and is an opportunity – even an obligation – to examine and seek to understand the wider system as a step towards influencing its direction. These papers demonstrate some of the breadth which will be required and some of the opportunities for positive change we could collectively choose to take.
Introducing a ‘Government as a System’ toolkit
It’s been fascinating to watch the iterative development of Policy Lab’s synthesis of how governments get things done. It has now mutated into something bigger and more ambitious, nothing less than a toolkit for developing and managing government as a system
Its centrepiece is a grid of 56 actions, mapped by approach to power and by position in the design cycle, There’s a huge amount of thought and experience baked into it, giving the potential to be a really valuable tool for framing issues in systems thinking terms.
From that perspective, there may be almost equal value in a second diagram, which expands the double diamond model into a chain of gems which map to the columns of the action grid.
But it’s important to recognise what this isn’t as well as what it is. It is a toolkit which is helpful in thinking about government as a system, it is not itself a depiction of that system. The grid is not a map, as the post at one point implies; rather it is a key to system maps as yet largely undrawn. An atlas of those maps, of various scales, complexity and precision would be a thing of wonder, but it is not an atlas we yet have – or probably could ever have, bringing to mind as it does Borges’ one paragraph short story, On Exactitude in Science. That’s not to diminish the power of the key and the approach, but it does very much reinforce the point that this is a toolkit, not a solution.
It will more interesting still to see where this might go next. This version is government as a system. The direction of travel points to a view which will increasingly be more about government in a system.
Taking a systems view of democratic design
Catherine Howe applies systems thinking to democracy in seven short sharp minutes. Don’t watch this for the production values – there aren’t any – but for a lot of thought neatly packed into a rather small space.
Democracy is
us figuring out as humans how to make the least worst decisions collectively
That definition combines extreme modesty with extraordinary ambitions, and very deliberately says nothing about democracy as representation. This is the broader, systemic view, built round three themes – networks of different scales (and so the network of those networks, and how people participate in them); design, of processes and structures; and the central importance of power and how it is distributed.
There’s no attempt to offer a solution. But better questions are always a good step towards better answers.
Computer Files Are Going Extinct
Files have been a unit of organisation for written information for generations. For a rather shorter period, they have been a unit of organisation for digital information. But after a few years in which the metaphor retained at least some connection with the underlying reality, computer files are abstracting themselves to vanishing point.
So what, who cares? Well at one level, none of us need care, that’s almost the point. At another level, those who have lived through the revolution in personalised computing can gently mourn – as the author of this post does – the loss of a way of organising and understanding information, and the sense of familiarity and control which comes with it. But as the author also recognises, all this starts to get really interesting when we recognise files are skeuomorphic (as is much of the language we use to talk about them).
But there is a more serious point too. However much we choose to abstract, in the end, the data is somewhere – and often in several somewheres. Nobody wants backups, as the old adage has it, but everybody wants restores. But the more our data becomes less apparent, less vulnerable to hard drive failures and laptop thefts, the more it becomes vulnerable to subscriptions not paid and services closed down.
This blog post has some characteristics which make it appear to be like a file. But it isn’t, it’s an entry in a database dressed up with some scripts. Its default state is not to exist, a state postponed by the application of money and electricity, but a state it will inexorably reach.
On being and doing in Government
Adrian Brown – Centre for Public Impact
What should government do? And, how should it do it? Those are two critically important questions, which fortunately get a lot of time and attention – even though it’s not hard to argue that they still don’t have good enough answers. But there is a third question which is at least as important, but which gets much less attention: what should governments be?
It is that question which is at the centre of this post. One reason why it has not had the attention it deserves is that a generation or two of public servants have been brought up not to notice it: the New Public Management paradigm that efficient delivery is pretty much all of what it’s about has become so pervasive as to be invisible. And that’s unfortunate in that it is neither value free (how could it be?) nor, as it turns out, is it a very good way of making governments work. NPM (and other strands of thought) are right that government does not exist for the benefit of people who work in it as politicians and officials. Its insights and methods have a place. But systems operated by and for humans need to have humans at their heart, and to recognise that it is the relationships and values those humans have which makes those systems work effectively – or even perhaps at all.
In which I am a bit over this digital transformation business.
It’s a pretty safe rule of thumb that whatever Catherine Howe is thinking about now, the rest of us will stumble onto at some point in the indefinite future. So if she is over the digital transformation business, we need to wonder where the zeitgeist will manifest next.
One of the more provocative definitions of technology is ‘everything which doesn’t work yet’. Similarly, we will know that mapping as a technique and transformation as a goal have become normal when we hardly need to talk about them, any more than we talk about the mature technology which is around us and so hardly needs to be spoken about. But that, as this post starts to explore, merely clears the ground for deeper and harder questions. The search is on for a theory of change to shape the search for answers.
The Moral Economy of Tech
This is another piece which isn’t new but which provides some good provocative food for thought, on how applying a computer programming perspective to problems which are fundamentally social can – and does – lead to unfortunate results. It’s written by Maciej Cegłowski, who brings elegant erudition to an unlikely range of subjects, in this case how an approach based on controlling closed systems breaks down when confronted with messily indeterminate systems, with a scattering of provocative one liners which combine challenge and simplicity, such as
Machine learning is like money laundering for bias.
But his conclusion is much broader – and an even greater challenge – than that single line suggests:
We have to stop treating computer technology as something unprecedented in human history. Not every year is Year Zero. This is not the first time an enthusiastic group of nerds has decided to treat the rest of the world as a science experiment. Earlier attempts to create a rationalist Utopia failed for interesting reasons, and since we bought those lessons at a great price, it would be a shame not to learn them.
Complex systems thinking is being used for policymaking. Is it the future?
Metaphors evolve. The fashion for mechanical metaphors to explain social phenomena is no longer as dominant as it once was (though such metaphors are still often lurking a little below the surface); the fashion for more organic metaphors is in the ascendant. Systems thinking generally and complex systems more particularly fit nicely with that trend. So it’s worth pausing to ask ourselves whether talking more about complex systems in the context of politics and policy making is merely following fashion or is getting closer to some underlying reality. Or to put that slightly differently, are complex systems an interesting metaphor for social and political systems or an accurate encapsulation of them?
This post gives some helpful pointers to answering that question, without quite actually answering it. Coincidentally, a Gordian knot-cutting tweet pops up from David Henig:
If a problem seems to be simple to solve, yet hasn’t been over many years, it’s probably complex.
One of the subheadings in the post asks ‘Is it really a new approach?’ The answer to that is clearly ‘no’ – the basic ideas have been known about in government for a long time. It’s tempting to get a bit recursive at that point: the reason why complex systems approaches have not become more deeply rooted in social and political change is itself an interesting complex systems problem.
Ten reform priorities for the new Prime Minister
New leaders rarely lack for advice on what their priorities should be and how they should approach them. This is a classic of its kind, well argued, well evidenced, addressing important issues – and yet missing something important in the gap between diagnosis and prescription.
The basic premise is that a political crisis should prompt a new prime minster to embrace structural reform of government, rather than to avoid or postpone it. That is almost certainly a forlorn hope – the capacity for reform of this kind is probably most available when the apparent need is least pressing – but that shouldn’t stop us reflecting on the merits of the ideas.
Many of the specific ideas put forward are sensible and serious, though there is a tendency to see centralisation and top down control as self-evidently ways of making things better. But the overall argument is undermined by missing out two big issues, both prompted by taking more of a systems perspective to the problem, which together point to the need for a theory of change to shape understanding of how real system improvement could be achieved.
The first is prompted by Stafford Beer’s aphorism, “the purpose of a system is what it does”. Observing that some aspects of the current do not work well and identifying alternatives which look as though they might work better is relatively easy. But it’s a safe assumption that nobody intended or wanted the system to work badly – the myth of civil service obstructiveness is exactly that – so to the extent that it does, understanding why the current system is as it is, and therefore whether different approaches would deliver different outcomes is less straightforward.
The second is that the system at issue is bigger than the one presented here and in particular that it is a political system. It is often tempting but often unhelpful to think of systems as machines, rather than as organisms, perhaps doubly so in political systems. Nobody should be criticised for wanting to change and improve things, but it is essential to recognise that if you want to change the system, you have to change the system.