Centre d’Etudes et de Recherches Appliquées à la Gestion

What is a Systemic Risk ?

The study of the Systemic Risk refers to the examination of the mechanisms, causes and consequences of severe financial crises. Because financial crises come in all sorts and flavors, providing a precise description and taxonomy of such events is a difficult task. It is commonly accepted to separate crises given their causes, and, especially, their consequences. For instance, a sudden crash affecting financial markets will be regarded as a financial shock given no default occur. Given such crash should lead to the failure of one, or more, financial institutions, we will refer to it as a financial crisis. Lastly, if the financial crisis is so severe that it affects the real economy (meaning economic growth, employment level, etc.), the crisis will be deemed systemic.
 
Here lies the complexity in assessing Systemic Risk. How to evaluate the possibility that a ‘regular’ financial crisis will turn into a systemic event?
 

Systemic Dimensions

The paper tries to provide an answer to this question. To do so, we introduce the concept of Systemic Dimensions. We pose the underlying assumption that Systemic Risk is in fact a composition of multiple/different risks. The assumption is relatively weak, as it seems difficult for a crisis to develop to such an extent without relying on various mechanisms, hence various risks. We consider three Systemic Dimensions: Losses, Illiquidity and Connectedness.
We provide a first validation that the Dimensions are actually non-correlated. In other words, that each Dimension does not assess a similar risk. We apply a Principal Component Analysis (PCA) that confirms the lack of correlation of Dimensions in stable periods (see Figure 1).
In a second time, we turn to the assessment of Systemic Risk using Systemic Dimensions. As Systemic is a combination of these Dimensions, a Systemic crisis should occur when all Dimensions increase. We propose a second (PCA), which highlights the fact that Systemic Dimensions are correlated in times of crisis.
 

Chronology in Systemic Events

There is however, a final point to address. Financial crises, as well as Systemic ones, unravel over a given period. Numerous phenomena occur one after the other, and, consequently, there should exists a specific chronology. In order to find evidences of such succession of events, we propose to lag (to move back in time) some Dimensions relatively to the others. By doing so, we expect to find stronger evidence of correlation in the Systemic Dimensions, hence validating our insight. Graphically, the lagging procedure is explained in Figure 2.
 
Empirically, we find evidences of such chronology. Connectedness tend to occur first, followed by Losses, while Illiquidity arises last. Our findings are robust over the U.S. financial markets, as well as in Europe. The study proposes a novel approach in defining, and assessing Systemic Risk.