Refine
Document Type
- Working Paper (12)
Language
- English (12)
Has Fulltext
- yes (12)
Is part of the Bibliography
- no (12)
Keywords
- Bank (5)
- Risikoverteilung (5)
- Schuldverschreibung (5)
- ambiguity (4)
- risk transfer (4)
- systematic risk (4)
- credit risk (3)
- laboratory experiments (2)
- ABS (1)
- Behavioral finance (1)
This paper analyzes the risk properties of typical asset-backed securities (ABS), like CDOs or MBS, relying on a model with both macroeconomic and idiosyncratic components. The examined properties include expected loss, loss given default, and macro factor dependencies. Using a two-dimensional loss decomposition as a new metric, the risk properties of individual ABS tranches can directly be compared to those of corporate bonds, within and across rating classes. By applying Monte Carlo Simulation, we find that the risk properties of ABS differ significantly and systematically from those of straight bonds with the same rating. In particular, loss given default, the sensitivities to macroeconomic risk, and model risk differ greatly between instruments. Our findings have implications for understanding the credit crisis and for policy making. On an economic level, our analysis suggests a new explanation for the observed rating inflation in structured finance markets during the pre-crisis period 2004-2007. On a policy level, our findings call for a termination of the 'one-size-fits-all' approach to the rating methodology for fixed income instruments, requiring an own rating methodology for structured finance instruments. JEL Classification: G21, G28
This paper analyzes the risk properties of typical asset-backed securities (ABS), like CDOs or MBS, relying on a model with both macroeconomic and idiosyncratic components. The examined properties include expected loss, loss given default, and macro factor dependencies. Using a two-dimensional loss decomposition as a new metric, the risk properties of individual ABS tranches can directly be compared to those of corporate bonds, within and across rating classes. By applying Monte Carlo Simulation, we find that the risk properties of ABS differ significantly and systematically from those of straight bonds with the same rating. In particular, loss given default, the sensitivities to macroeconomic risk, and model risk differ greatly between instruments. Our findings have implications for understanding the credit crisis and for policy making. On an economic level, our analysis suggests a new explanation for the observed rating inflation in structured finance markets during the pre-crisis period 2004-2007. On a policy level, our findings call for a termination of the 'one-size-fits-all' approach to the rating methodology for fixed income instruments, requiring an own rating methodology for structured finance instruments. JEL Classification: G21, G28 Keywords: credit risk, risk transfer, systematic risk
Risk transfer with CDOs
(2008)
Modern bank management comprises both classical lending business and transfer of asset risk to capital markets through securitization. Sound knowledge of the risks involved in securitization transactions is a prerequisite for solid risk management. This paper aims to resolve a part of the opaqueness surrounding credit-risk allocation to tranches that represent claims of different seniority on a reference portfolio. In particular, this paper analyzes the allocation of credit risk to different tranches of a CDO transaction when the underlying asset returns are driven by a common macro factor and an idiosyncratic component. Junior and senior tranches are found to be nearly orthogonal, motivating a search for the whereabout of systematic risk in CDO transactions. We propose a metric for capturing the allocation of systematic risk to tranches. First, in contrast to a widely-held claim, we show that (extreme) tail risk in standard CDO transactions is held by all tranches. While junior tranches take on all types of systematic risk, senior tranches take on almost no non-tail risk. This is in stark contrast to an untranched bond portfolio of the same rating quality, which on average suffers substantial losses for all realizations of the macro factor. Second, given tranching, a shock to the risk of the underlying asset portfolio (e.g. a rise in asset correlation or in mean portfolio loss) has the strongest impact, in relative terms, on the exposure of senior tranche CDO-investors. Our findings can be used to explain major stylized facts observed in credit markets.
Risk transfer with CDOs
(2008)
Modern bank management comprises both classical lending business and transfer of asset risk to capital markets through securitization. Sound knowledge of the risks involved in securitization transactions is a prerequisite for solid risk management. This paper aims to resolve a part of the opaqueness surrounding credit-risk allocation to tranches that represent claims of different seniority on a reference portfolio. In particular, this paper analyzes the allocation of credit risk to different tranches of a CDO transaction when the underlying asset returns are driven by a common macro factor and an idiosyncratic component. Junior and senior tranches are found to be nearly orthogonal, motivating a search for the where about of systematic risk in CDO transactions. We propose a metric for capturing the allocation of systematic risk to tranches. First, in contrast to a widely-held claim, we show that (extreme) tail risk in standard CDO transactions is held by all tranches. While junior tranches take on all types of systematic risk, senior tranches take on almost no non-tail risk. This is in stark contrast to an untranched bond portfolio of the same rating quality, which on average suffers substantial losses for all realizations of the macro factor. Second, given tranching, a shock to the risk of the underlying asset portfolio (e.g. a rise in asset correlation or in mean portfolio loss) has the strongest impact, in relative terms, on the exposure of senior tranche CDO-investors. Our findings can be used to explain major stylized facts observed in credit markets.
Decisions under ambiguity depend on both the belief regarding possible scenarios and the attitude towards ambiguity. This paper exclusively investigates the belief formation and belief updating process under ambiguity, using laboratory experiments. The results show that half of the subjects tend to adopt a simple heuristic strategy when updating beliefs, while the other half seems to partially adopt the Bayesian updates. We recover beliefs, represented by distributions of the priors/posteriors. The recoverable initial priors mostly follow a uniform distribution. We also find that subjects on average demonstrate slight pessimism in an ambiguous environment.
This paper provides a systematic analysis of individual attitudes towards ambiguity, based on laboratory experiments. The design of the analysis allows to capture individual behavior across various levels of ambiguity, ranging from low to high. Attitudes towards risk and attitudes towards ambiguity are disentangled, providing pure measures of ambiguity aversion. Ambiguity aversion is captured in several ways, i.e. as a discount factor net of a risk premium, and as an estimated parameter in a generalized utility function. We find that ambiguity aversion varies across individuals, and with the level of ambiguity, being most prominent for intermediate levels. Around one third of subjects show no aversion, one third show maximum aversion, and one third show intermediate levels of ambiguity aversion, while there is almost no ambiguity seeking. While most theoretical work on ambiguity builds on maxmin expected utility, our results provide evidence that MEU does not adequately capture individual attitudes towards ambiguity for the majority of individuals. Instead, our results support models that allow for intermediate levels of ambiguity aversion. Moreover, we find risk aversion to be statistically unrelated to ambiguity aversion on average. Taken together, the results support the view that ambiguity is an important and distinct argument in decision making under uncertainty.
This paper applies the theory of structured finance to the regulation of asset backed securities. We find the current regulation in Europe (Article 405 of the CRR) and the US (Section D of Dodd-Frank Act) to be severely flawed with respect to its key intention: the imposition of a strict loss retention requirement. While nominal retention is always 5%, the true level of loss retention varies across available retention options between zero loss retention and full loss retention at the extreme ends. Based on a standard model of structured finance transactions, we propose a new risk retention metric RM measuring the level of an issuer’s skin-in-the-game. The new metric could help to achieve a better implementation of CRR/CRD-IV and DFA, by making disclosure of the RM-number compulsory for all ABS transactions. There are also implications for the operation of rating agencies. On a general level, the RM metric will be instrumental in achieving simplicity and transparency in securitizations (STS).
We develop a state-space model to decompose bid and ask quotes of CDS into two components, fair default premium and liquidity premium. This approach gives a better estimate of the default premium than mid quotes, and it allows to disentangle and compare the liquidity premium earned by the protection buyer and the protection seller. In contrast to other studies, our model is structurally much simpler, while it also allows for correlation between liquidity and default premia, as supported by empirical evidence. The model is implemented and applied to a large data set of 118 CDS for a period ranging from 2004 to 2010. The model-generated output variables are analyzed in a difference-in-difference framework to determine how the default premium, as well as the liquidity premium of protection buyers and sellers, evolved during different periods of the financial crisis and to which extent they differ for financial institutions compared to non-financials.
The pricing of an ambiguous asset, whose cash flow stream is uncertain, may be affected by three factors: the belief regarding the realization likelihood of cash flows, the subjective attitude towards risk, and the attitude towards ambiguity. While previous literature looks at the total price discount under ambiguity, this paper investigates with laboratory experiments how much effect each factor can induce. We apply both non-parametric and parametric methods to cleanly separate the belief effects, the risk premiums, and the ambiguity premiums from each other. Both methods lead to similar results: Overall, subjects have substantial ambiguity aversion, and ambiguity premiums account for the largest price deviation component when the degree of ambiguity is high. As information accumulates, ambiguity premiums decrease. We also find that beliefs do influence prices under ambiguity. This is not because beliefs are biased towards either good or bad scenarios per se, but because subjects display sticky belief updating as new information becomes available. The clear separation performed in this paper between belief and attitude also enables a more accurate estimation of the parameter of ambiguity aversion compared to previous studies, since the effect of beliefs is partialled out. Overall, we find empirically that both factors, belief and attitude towards ambiguity, are important factors in pricing under ambiguity.
This paper sets up an experimental asset market in the laboratory to investigate the effects of ambiguity on price formation and trading behavior in financial markets. The obtained trading data is used to analyze the effect of ambiguity on various market outcomes (the price level, volatility, trading activity, market liquidity, and the degree of speculative trading) and to test the quality of popular empirical market-based measures for the degree of ambiguity. We find that ambiguity decreases market prices and trading activity; ambiguity leads to lower market liquidity through wider bid-ask spreads; and ambiguity leads to less speculative trading. We also find that popular market-based measures of ambiguity used in the empirical literature do not seem to correctly capture the true degree of ambiguity.