Refine
Year of publication
- 2014 (3) (remove)
Document Type
- Working Paper (3)
Language
- English (3)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Keywords
- Systemic risk (3) (remove)
Institute
- Center for Financial Studies (CFS) (3) (remove)
In this paper, we propose a novel approach on how to estimate systemic risk and identify its key determinants. For all US financial companies with publicly traded equity options, we extract their option-implied value-at-risks (VaRs) and measure the spillover effects between individual company VaRs and the option-implied VaR of an US financial index. First, we study the spillover effect of increasing company risks on the financial sector. Second, we analyze which companies are most affected if the tail risk of the financial sector increases. We find that key accounting and market valuation metrics such as size, leverage, balance sheet composition, market-to-book ratio and earnings have a significant influence on the systemic risk profile of a financial institution. In contrast to earlier studies, the employed panel vector autoregression (PVAR) estimator allows for a causal interpretation of the results.
We develop a methodology to identify and rank “systemically important financial institutions” (SIFIs). Our approach is consistent with that followed by the Financial Stability Board (FSB) but, unlike the latter, it is free of judgment and it is based entirely on publicly available data, thus filling the gap between the official views of the regulator and those that market participants can form with their own information set. We apply the methodology to annual data on three samples of banks (global, EU and euro area) for the years 2007-2012. We examine the evolution of the SIFIs over time and document the shifs in the relative weights of the major geographic areas. We also discuss the implication of the 2013 update of the identification methodology proposed by the FSB.
We study self- and cross-excitation of shocks in the Eurozone sovereign CDS market. We adopt a multivariate setting with credit default intensities driven by mutually exciting jump processes, to capture the salient features observed in the data, in particular, the clustering of high default probabilities both in time (over days) and in space (across countries). The feedback between jump events and the intensity of these jumps is the key element of the model. We derive closed-form formulae for CDS prices, and estimate the model by matching theoretical prices to their empirical counterparts. We find evidence of self-excitation and asymmetric cross-excitation. Using impulse-response analysis, we assess the impact of shocks and a potential policy intervention not just on a single country under scrutiny but also, through the effect on cross-excitation risk which generates systemic sovereign risk, on other interconnected countries.