The Financial Risk Meter (FRM) helps you to identify different systemic risk level in the financial market over time. It is an index of the system volatility level which indicates that if FRM is high, then the systemic risk is high.
The blue time series shows the FRM over its entire historical range for the given market. Below it (for available markets) is a granular time series illustration of FRM data from a shorter, selected range in recent or immediate history. For each date, the risk values of each constituent financial body are represented as a box plot. Notice how the red curve, tracing the maximums at each time step, leads the blue curve in upward and downward trends – that is to say, the most vulnerable nodes of the network which are affected most from systemic risk act as early indicators for subsequent movement of other nodes in the cohort.
We propose a linear lasso measure to estimate systemic interconnectedness across financial institutions based on tail-driven spill-over effects in an ultra-high dimensional framework. Methodologically, we employ a variable selection technique in a time series setting for a linear quantile regression framework with 5% quantile. We can thus include more financial institutions into the analysis, to measure their interdependencies in tails.
Then FRM is induced from this model which is the averaged tuning parameter lambda from lasso technique. The estimation method of it is cross validation. In application we apply 100 US publicly traded financial institutions and 6 macro state variables to estimate this index. Previously we have used 200 financial institutions, after comparison we find out that using 100 firms is more efficient way.
The recession indicators for the US and Euro Area can be found at the St. Louis Fed and CEPR websites, both using a similar methodology. According to the NBER, a recession is a significant decline in economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales. The recession indicator has a value of 1 in a recessionary period, while a value of 0 indicates an expansionary period.
In the paper we detail how the FRM shows a predictability to an imminent recession and serves therefore as an indicator for systemic risk in a variety of world regions. We, therefore, suggest that FRM can be considered in the inclusion of the list of leading indicators and is informative in terms of predicting upcoming recessions.
We provide here a subset of the FRM time-series for public download. This sample data is from the previous 30 days since the most recent update to the market's index. With this sample you can get a first look into the FRM numerically in addition to the visual representation in the time series above.
Red: severe risk of a crisis in the financial market. Our risk measure suggests that a financial crisis is imminent or happening right now. This risk level is given for lambda values higher than the 80%-ratio.
Orange: high risk of crisis in the financial market. A crisis might occur very soon. This risk level is given for lambda values between the 60%-ratio and 80%-ratio.
Yellow: elevated risk of crisis in the financial market. The incidence of a crisis is somewhat higher than usual. This risk level is given for lambda values between the 40%-ratio and 60%-ratio.
Blue: general risk of crisis in the financial market. There is no specific risk of a crisis. This risk level is given for lambda values between the 20%-ratio and 40%-ratio.
Green: low risk of crisis in the financial market. The incidence of a crisis is less likely than usual. This risk level is given for lambda values lower then the 20%-ratio.
FRM Financial Risk Meter (2020)
Advances in Econometrics, The Econometrics of Networks, vol. 42
Mihoci A, Althof M, Chen CYH, Härdle WK
An AI approach to Measuring Financial Risk (2020)
Singapore Economic Review
Yu L, Härdle WK, Borke L, Benschop T
LASSO-Driven Inference in Time and Space (2018)
IRTG 1792 Discussion Paper 2018-021
Victor Chernoyhukov, Wolfgang Karl Härdle, Chen Huang, Weining Wang
Single-Index-Based CoVaR With Very High-Dimensional Covariates (2017)
Journal of Business&Economic Statistics
Yan Fan, Wolfgang Karl Härdle, Weining Wang and Lixing Zhu
TENET: Tail-Event driven NETwork risk (2016)
Journal of Econometrics
Wolfgang Karl Härdle, Weining Wang, Lining Yu
QuantNet is designed as a web-interface to freely exchange numerical methods, called Quantlets