Why we need network analysis to understand the future of economics

Picture by AP/Press Association Images

Picture by AP/Press Association Images

What it is, what it’s not

Network analysis is the method of the future. That is not only – certainly not primarily – because we are ever more connected in some superficial social-media driven internet sort of way. All of that may be fascinating (and certainly can be analysed using network analysis), but it is not fundamental to our existence as humans – we existed before Facebook, we will exist after it is gone.

Entirely fundamental though are the complex linkages between humans, problems and resources. And those linkages are just as important as the humans, problems and resources themselves. Analysing the links, not just the elements in isolation, requires network analysis.


The problem

In environmental, human and, therefore, long-run economic terms the models we use to describe the world currently find false optimal flight-paths towards unsustainable monolithic solutions. And don’t forget what an important and multi-faceted word unsustainable is – not just environmental concerns, but also the physical and mental health of populations, poverty and income divergence, political and societal fractures.

Human society is ever more linked. But the business, wider economic and political imperative hangs doggedly onto an assumption of individualism. And alongside this grand assumption sit the linear, non-network methods of analysis. It is hard to say which way cause or effect works – almost certainly some in both directions. And anyway, these traditional ways of seeing the world produce apparent ‘knowledge’ (or, even more dangerously,’solutions’) whilst in fact pushing the societal direction of travel entirely the wrong way.


Networked animals

Embedded within the definition of network analysis is its proximity to our human experience. Network data occurs whenever there is

  • some kind of ‘entity’, be that a human agent, an event, a geographical location, and
  • some kind of linkage or relationship between these e.g. humans meeting, events of a similar nature or occurring simultaneously, geographical places linked by transport.

A simple example of how this contrasts to non-network analysis is on risks of communicable disease. A non-network model would assign the risk of disease to someone according to characteristics: where they live, their income level, general health status, etc. But if we bring in the power of networks, understanding who had a relationship with who, we can analyse how someone is positioned in the network. If they are where many people had the disease and links were many and strong, or if very close to several people who were at high risk then that would indicate a high risk of contracting the disease. Clearly, with networks included we build a much more powerful model.


A better world…

Co-operation was shown many years ago to be the optimal solution in a vast range of situations, far outperforming the blind pursuit of individual interest. But this fact is ignored by most of the human systems that are shaped and built by government and business. In exactly the same way the reality of a connected world is ignored in decision-making models from big data, through HR ‘performance systems’, health, education and other metrics, GDP and other economic statistics. Ultimately, we have to understand linkages and feedback in network models and reform our thinking all the way to the classic (linear) economic model where, most dangerously, the assumption of ‘independence’ is so heavily embedded it cannot be escaped.


…based around humans and the planet

Dynamic, interconnected analysis approaches built around networks (and associated complexity methods) are able to create more human-centred  and sustainable directions – also they can reveal the weaknesses in our society built on an ignorance of complexity. If we model who we really are, what we really do and our relationship with a complex world more faithfully and subtly we can make progress. Such models illustrate the potential of shifting and changing solutions rather than a distracting and damaging simple point-estimate.


“Models are opinions embedded in mathematics”

Perhaps solidarity and co-operation have gone out of fashion. Perhaps an empathy with the natural world is ebbing away. Or maybe these values stand no chance in a world shaped around the flawed machine algorithms and models that now measure and decide our lives.

Some models don’t ignore this, such as many trading algorithms for financial instruments, and they succeed greatly – in a sense – by making large profits for those who run them and dumping the costs on us. Partly because our models don’t recognise what theirs do.

So the knowledge is out there, but not being used for our benefit, yet! We should demand better in the models that shape our everyday lives – and follow the best. We must adopt network analysis widely to embrace concepts which model our modern human reality and reject the outdated, disconnected and linear view of the world.

  • Good article, and very true. A bit worried that “network analysis” might become hypostasised into yet another empty buzzword by those who can’t be bothered to understand it. This mode of analysis goes by other names too, like “systems thinking”, and is already practised effectively by, for example, Vaclav Smil or Jeffrey Sachs.

  • Gustav Clark

    Very good indeed. You are right that this is not new, and also right that it is the way forward. The reason that we could not manage it in the past was that we simply didn’t have the analytical power. Economics was forced to use the old linear methods because they just could not handle the full complexity of the world, and its ineffectiveness made it largely the preserve of universities and government. I compare it to automatic speech recognition – for years it failed to deliver because all our attempts were based on the implementation of theoretical linguistic models. Our current success came about because companies like Google have such massive computing power that they can learn the patterns. No one in the economics world has that technical mindset coupled to that computing power and so we are still struggling.
    However, in a sense a networked approach is part of the present. There is a common assumption that economics failed in the banking crisis, but it is not a monolithic science. There are may people working on the whole interrelatedness of the world economy, the crash made their work more important. The problem is how to ensure that they are funded, with an acceptance that failure is likely and that success won’t be overnight. That is of course the classic context where we hope that private finance will step in. If I were an investment banker I would demand the best tools for the job, and be prepared to see some of my money funding a few startups in the hope that one of them will deliver. The pressure there is much greater than any academic or government economist will ever feel and many will fail. But with luck a few will succeed and be vindicated by the increased profits of those who use the tools. That will provide the justification for academia and government to join in.
    I’ve developed my ideas on this whilst writing, but it strikes me that there may be a general truth in the idea that only private finance can accept the risks implicit in intellectual work where there is no guaranteed reward. Fintech did look like a promising new industry for Britain – I hope Mrs May doesn’t stifle it.

openDemocracyUK presents a debate about how to build a just, sustainable, and resilient economy. Find out more about the project & submit your ideas.

  • All
  • Civil Society
  • Constitution
  • Education
  • Infrastructure
  • Local Government
  • Measurement
  • Money
  • Ownership
  • Procurement
  • Regulation
  • Research and Development
  • Spending
  • Tax
  • Trade policy