Systemic risk & catataxic shift


Specific vs Systemic risk

One of the most common ways to reduce risk is to diversify. In other words, as the saying goes:

“Don’t put all your eggs in one basket”

But this simple motto masks a more profound reality which is this: not all types of risk can actually be reduced by diversification. The important distinction to make is between systemic risk and specific risk. Specific risk can be reduced by diversification. Systemic risk can not. 

What do we mean by the terms ‘specific’ and ‘systemic’? Specific risk is a discrete type of risk that only applies to a single entity whereas systemic risk is a risk that applies to all entities in the system. Imagine you are a property investor and own a large number of buildings in a city. You specific risk is the risk that one of the buildings in your property portfolio burns down in a fire. Your systemic risk is the risk that the whole city is destroyed in an earthquake. 

We can express this using the eggs and baskets metaphor. The breaking of a single egg is the specific risk. The destruction of the whole basket represents the systemic risk. The problem comes when you think you have put your eggs in different baskets only to find out they are actually all in the same (bigger) basket.

Diversification failure and the Phase shift

One of the best conceptual analogies to use here is the phase shift: when a substance changes from solid to liquid and then to gas. Consider a gas first. All the molecules in a gas act independently of each other. If you were to release a gas in an enclosed space  (maybe a contented belch after lunch?) it would gradually diffuse throughout the room. The way one particle moves only affects another when they bump into each other. These random collisions are the reason a gas will evenly spread to fill the space available. But, other than in collisions, you can consider each gas molecule as an independent entity following its own random path. 

As the temperature changes and the gas becomes liquid, this is no longer the case. In a liquid, the molecules begin to have an influence over each other creating a more coherent flow of fluid. So we can say that the molecules in a fluid have a greater degree of interconnectivity than in a gas. As the temperature drops again and the substance freezes, each molecule is locked into a rigid crystalline structure; they have no independent movement at all. It can no longer be considered as a collection of independent molecules but rather as single collective entity – an ice cube. 

In this example we have a variable – temperature- which dictates whether you should be viewing the substance as a single monolithic entity or as a a collection of individual molecules. This shift in viewpoint, from the parts to the whole, is known as the catataxic shift. So we can say that, in this example, temperature drives the catataxic shift by changing the degree of interconnectivity in the system.  

Let’s take another example and look at traffic on a motorway. Imagine you are driving on a relatively empty stretch of road with few other vehicles. You are pretty much free to drive as you like. There is so much space to manoevre  that you don’t need to pay much attention  to what other drivers are doing. But as the motorway becomes more crowded you need to start paying more attention. Your driving involves reacting to what other drivers are doing to a much greater extent. When they break, you break. When a lorry pulls out to overtake another, you are forced to slow down. Then, of course, we reach a critical point when the traffic stops completely – a traffic jam. 

These three phases are just like the gas-liquid-solid phases earlier. The empty road is like the gas phase; cars moving independently of each other. The traffic jam is like the solid phase; the cars are no longer individual entities but have formed a single monolithic block. In this example, the number of cars on the road (traffic density) drives the Catataxic shift.

The catatactic shift – buttons and threads

The mathematics behind this sudden phase shift is part of network theory and can be explained using the example of button and threads. Imagine scattering 20 buttons on a table and then handing out a thread to a volunteer and asking them to connect two random buttons together with the thread.  You then pick up that thread. Clearly, you will also then pick up the two buttons as well. As you continue connecting the threads and buttons together at random there will come a time when if you pick up a thread you will pick up all the buttons on the table. The question is when will this happen?

The answer is surprising: it happens earlier and much more suddenly than you might think. It happens when the ratio of threads to buttons approaches 1 to 2.  In other words for 20 buttons, once you have handed the volunteer 10 threads, picking up a single thread will pick everything up. Look at the chart above. You can see a sudden dramatic shift  – the Catataxic shift – when the threads to buttons ratio hits 0.5. This is the equivalent of water suddenly freezing in a river or a traffic jam suddenly appearing out of nowhere on the motorway.

The Catataxic shift is the key reason why diversification strategies fail. Remember that specific risk can be diversified but systemic risk can not. When things that you thought were unconnected suddenly become interconnected you introduce systemic risk. That means that the key question for diversification strategies is not “how correlated are these things now?” because if the correlation were visible they would not be diversification candidates. Instead, the key question is “what factor could drive a Catataxic phase shift?”. What is the equivalent of temperature or traffic density in my environment that will cause the whole system to seize up?