What Matters: The networked market

Previous: A bigger tent for financial reform

The networked market

23 February 2009

Many of the world’s economic leaders, from US Federal Reserve Chairman Ben Bernanke on, have been surprised by the suddenness, intensity, and duration of the current global economic storm. It is clear that something about the models they have been using, both explicitly to forecast the future and implicitly to guide their policy responses, have led them astray. What was it?

Lying at the heart of the most common market theory—the efficient-market hypothesis—is the assumption that financial returns, including stock prices, are independent of one another. The mathematical signature of this assumption is what statisticians call the IID assumption, where stock price changes are independent of each other and drawn from identical distributions. Under these assumptions, stock price changes are randomly or log-normally distributed.

But we now know without any doubt that this assumption is wrong. In fact, stock price changes more closely match a “fat tail” distribution rather than a random distribution. In a fat-tail, or “power law,” distribution, there are many more very large changes (either positive or negative) than in a random distribution.

Random distributions come from collections of independent components not connected to one another. Power laws come from “networks” in which the individual components are very much connected to one another in a (potentially vast) system of nodes and links (for example, the price of Goldman Sachs’s stock today is not independent of the price changes in Morgan Stanley’s stock yesterday).

Why is this important? Because models based on an assembly of random components cannot result in “shocks”—outcomes outside the range of conventional expectations. Change happens in a very regular way in models based on independent components.

This is not at all the case with networks, and particularly networks with feedback loops (for instance, low stock prices result in lower savings for investors, which results in lower purchases, which results in lower earnings for companies, which again lowers stock prices, ad nauseam) or feed-forward loops based on often erroneous forecasts. Shocks—cascading failures of large sections of the system—can arise naturally in a network system, though one does not always know exactly where or when they will occur. Shocks occur as a direct consequence of the network’s design. Accordingly, if we are going to make progress in anticipating and dealing with shocks, we need to jettison the assumption of randomness and begin to model our financial systems for what they really are: networks. If we want to reduce the frequency or impact of the shocks, we have to rethink and redesign our financial networks to minimize shocks.

These models may take many years to get right. But we should start down that path. It has taken us decades to model global weather, but now we get invaluable early warning of coming storms. We need better mathematical models of our financial systems as well to give us early warning of economic storms. Too much is at stake to avoid the challenge.

Comment [2]

Agree? Disagree? Let us know what you think. Please include your full name with your comment. Comments may be edited.

One way to manage risk is not to try to predict and avoid, but to build a system that is resilient. There are several priciples to building system resiliency, two of which are (1) loose, rather than tight, coupling and (2) diversity of elements. It’s interesting (and perhaps important) to speculate about whether a globally networked economy, with tight and almost instantaneous connections between geographically specialized components, might be subject to types of catstrophic failures than the reverbarations of current credit crisis.

Posted 15 June 2009, 11:20 by Richard Christner
I agree that that to find models that will get it “right: will take many years.
Having said this the current collapse was totally predictable based on the underlying system dynamics, What was not predictable was the timing of the collapse- delusion and wishful thinking are more powerful that logic. in other words the structure of the current economic system guaranteed its collapse – a system that requires debt to grow faster than the income stream that supports it is doomed fail at some point. The asset price bubble that was required to support the this growth debt was mistaken as a sustainable reality is now slowly being recognised as a delusion by those who recognise that “green shoots” are a manufactured attempt at hoping that positive market sentiment can overcome reality in the long run.
Having said this, traders focus on the short term and momentum. While they may rationally or intuitively understand that endless debt and economic growth on a limited planet is unsustainable, both the logic of the system and the feedback systems mean that they have to ignore long term reality for the short term sentiments of the market if they are to survive.

In my opinion, the only way we are going to find the models you are talking about is to create an environment that rewards the behaviour that will deliver this outcome. Unfortunately the realities of politics is that it has up to now had to focus on the short term – the creation of jobs and economic growth in the short term even if this guarantees longer term failure are a prerogative – especially when there is global collusion by world leaders that this is required .

My opinion is that we have finally hit the wall to economic growth. We dodged this bullet in 1971 by leaving the gold standard. This gave us the opportunity to believe we could get rich and grow the economy by pretending that our existing assets were becoming more valuable because we could sell them amongst ourselves for n ever increasing price (due in part to falling interest rates) .

I don’t know what more to say other than the reality of our predicament has not yet played itself out fully. In Australia it still looks like we have escaped. But this is a myth. The collapse in demand started in the USA. The next cabs off the rank were the counties that relied on exports to the USA – Germany, Japan and China. The last to feel these effects are the countries that supply the raw material to these countries – eg Australia.

Moreover, the estimated $4.5T losses of capital by banks by the IMF will shrink the global debt available by $45T due to the realities of the way that banks can lend. (reserve Ratio) This has yet to have it’s full effect on the global economy.. When it does many companies and countries will find that they are unable to roll over the debt that they require to function – especially as the value of the assets they used to justify their borrowing falls. This will induce “second order” effects ….

I find watching this process of slow collapse an amazing learning process. World leaders and reserve banks are either in denial or are acting on an unattainable wing and a prayer. With debt to GDP ratios is in excess of twice what they were in 1929 in the USA and other developed countries, combined with the realities of structural global energy and general commodity shortages, I find it difficult to find the math or delusion that will pull us out of this with a return to growth scenario

Posted 7 June 2009, 08:12 by Alan McCrindle

Commenting is closed for this article.

Previous: A bigger tent for financial reform

Next: It’s the leverage, stupid

Send an e-mail to let us know how we can make our site better.

A better way to fix the banks
McKinsey directors Lowell Bryan and Toos Daruvala present a plan that could solve the toxic-asset pricing problem voluntarily—without requiring Uncle Sam to nationalize the whole industry—and make (pretty much) everyone a winner. (The McKinsey Quarterly)
Lessons of the Financial Crisis
CFR Senior Fellow and Director of International Economics Benn Steil has provided an incisive explanation of the global financial crisis and its causes and a set of practical reforms to prevent its recurrence.
McKinsey on the new role of government
McKinsey's Lenny Mendonca discusses on Big Think, a global online forum, the challenges and opportunities emerging within a new global regulatory environment.