A liquidity risk stresstesting framework with interaction between Market and Credit Risk.docx

1、A liquidity risk stresstesting framework with interaction between Market and Credit Risk1 A liquidity risk stress-testing framework with interaction between market and credit risks Eric Wong*+ and Cho-Hoi Hui* This version: April 11, 2011 Abstract This study develops a framework for stress testing b

2、anks liquidity risk, where liquidity and default risks can stem from market risk arising from asset price shocks. The risks are assumed to be transmitted through three channels. First, banks mark-to-market losses on assets increase their default risk and thus induce deposit outflows. Second, asset p

3、rice declines reduce banks asset liquidity. Third, banks also face higher contingent liquidity risk, as the likelihood of drawdowns on irrevocable commitments by customers increases in the stress scenario. Contagion risk is also incorporated in the framework through banks linkages in the interbank a

4、nd capital markets. The framework quantifies liquidity risk by estimating the expected cash shortage time and the expected default time of banks. Stress testing results from this framework suggest that liquidity risk in the Hong Kong banking sector would be contained even if such stress scenario wer

5、e to occur. JEL classifications: C60, G21, G28 Key words: Liquidity risk, stress testing, default risk, market risk, banks, Hong Kong * Research Department, Hong Kong Monetary Authority, 55/F, Two International Finance Centre, 8, Finance Street, Central, Hong Kong, China. Email: etcwonghkma.gov.hk,

6、and chhuihkma.gov.hk Phone: (852) 2878 8735 Fax: (852) 2878 1891 + Correspondence. The authors would like to thank Naohiko Baba, Klaus Duellmann, Hans Genberg, Martin Grieder, participants at the third annual workshop of the Asian Research Networks on 25 March 2010 co-organised by the Bank for Inter

7、national Settlements and Bank of Japan; the research workshop “Challenges in Banking Research” on 28-29 May 2009 organised by the Research Task Force of the Basel Committee on Banking Supervision and sponsored the Bank of Spain, and “the Third Stress Testing Expert Forum: Advanced Techniques in Stre

8、ss Testing” on 19-20 May 2009 organised by the IMF and Deutsche Bundesbank for their useful suggestions and comments. The views and analysis expressed in the paper are those of the authors, and do not necessarily represent the views of the Hong Kong Monetary Authority.1 1. Introduction As illustrate

9、d by recent developments in the US and European banking systems, liquidity and default risks of banks can stem from the crystallisation of market risk and such interaction of risks could lead to systemic crises, e.g. the sub-prime crisis emerged in 2007.1 While the banking systems in most other econ

10、omies have shown relatively resilient, they are not immune to similar crises because of three common features running through all banking systems. First, banks balance sheets are inevitably exposed to common market risk factors, as they generally hold similar financial assets. Thus, significant asse

11、t price declines, even only in a single market, could expose a large number of banks to substantial market risk losses. Secondly, the capital available for banks to buffer against such market risk losses is limited, as banks usually operate with a high level of financial leverage. This suggests that

12、 all banking systems are essentially vulnerable to multiple default risk during severe market risk shocks. Thirdly, interbank markets are extremely sensitive to default risk. Any shock that increases the default risk of banks could result in tightened interbank markets, creating systemic liquidity s

13、hortages. For banking stability it is, therefore, important to assess the extent to which a banking system is exposed to such an interaction of risks. However, in the literature, stress-testing frameworks capturing the interaction of risks are relatively scant. To fill this gap, this study develops

14、a new stress-testing framework to assess the liquidity 1 The development of the sub-prime crisis can be summarised as follows: In the early stage (around the second half of 2007), there were continual announcements of significant mark-to-market write-downs of sub-prime mortgage-related securities by

15、 financial institutions in the US and Europe as a result of deterioration in the asset quality of sub-prime mortgages. Such crystallisation of market risk triggered concerns of banks default risk quickly, as evidenced by increases in credit default swap spreads of banks since the third quarter of 20

16、07. Default risk of banks was amplified further following the debacle of some large financial institutions, including Bear Stearns, Freddie Mac and Fannie Mae. As default risk heightened, banks became increasingly reluctant to lend among themselves, resulting in the systemic liquidity problems in th

17、e global interbank markets in mid-September 2008 following the collapse of Lehman Brothers, despite the unprecedented actions and measures taken before by various central banks and governments to inject liquidity in the global banking system. 2 risk of banks in this context. In the framework, we ass

18、ume that there is a prolonged period (i.e. one year) of negative exogenous asset price shocks in some major financial markets, which affect banks liquidity risk through three channels: (i) increases in banks default risk and deposit outflows; (ii) reduction in banks liquidity generation capability;

19、and (iii) increases in contingent drawdowns. Default risk of banks is endogenously determined using a Merton-type model in the framework.2 Contagious default risk is incorporated through banks linkage on interbank and capital markets that is consistent with the theories in the literature. With this

20、framework, daily cash outflows of banks can be simulated given exogenous asset price shocks. Using the Monte Carlo method, the framework quantifies the liquidity risk of individual banks by estimating the probability of cash shortage and the probability of default due to liquidity problems. In addit

21、ion, conditional on occurrences of cash shortage and default in the simulations, the first cash shortage time and the default time can be estimated respectively. The corresponding probability of multiple defaults of banks in a banking system can be also estimated, which is an important measure for a

22、ssessing the systemic risk in the banking system. The framework with two stress scenarios is applied to assess the liquidity risk of a group of 12 listed banks in Hong Kong with publicly available data. This study draws on the literature that relates to the impact of asset price 2 Using Merton-type

23、models to endogenise default risk of banks in systemic risk assessment frameworks is also adopted by Aspachs et al. (2006), which is based on the theoretical framework by Goodhart et al. (2006). 3 declines on banks default risk. Cifuentes et al. (2005) and Adrian and Shin (2008a, 2008b) provide a th

24、eoretical foundation on how a small asset price shock can be amplified by its mark-to-market (MTM) effects on banks balance sheets, and thus leads to a downward spiral in asset prices and contagious defaults of banks through interbank linkages.3 The importance of the linkage between asset prices and

25、 default risk of banks has recently been recognised and studied empirically by Boss et al. (2006) and Alessandri et al. (2007). However, the implications for banks liquidity risk, which are crucial for policy markers in view of the sub-prime crisis, are not the main focus of these studies. van den E

26、nd (2008) incorporates the interaction between market and funding liquidity and potential feedback on banks into a framework to assess liquidity risk of Dutch banks by estimating the distributions of liquidity buffers and probability of liquidity shortfall of banks. However, the effect of market ris

27、k shocks on default risk of banks and in turn their deposit outflows (and liquidity risk) are not explicitly modelled in the framework. By contrast, the proposed framework in this paper tries to establish the linkages between liquidity risk, market risk and banks default risk. It is also related to

28、the theoretical framework by Diamond and Rajan (2005) which shows that an aggregate liquidity shortage can be caused by bank runs perceived by bank insolvency. This study contributes to the literature in two aspects. First, this is among few empirical studies to incorporate interaction of risks in a

29、 liquidity risk stress-testing framework. Given that the sub-prime crisis is highly relevant to such interaction of risks, the framework could be useful for policy makers to assess how resilient a banking sector is under liquidity shocks similar to or even severer than 3 See also Allen and Gale (199

30、4, 1998 and 2000a), Acharya and Yorulmazer (2007), and Nier et al (2007). In addition to the MTM effects, asset price collapse could cause bank failures because of widespread defaults of banks lending to investors for acquiring risky assets (see Allen and Gale (2000b) and Goetz Von Peter (2004). 4 t

31、hose occurred in the sub-prime crisis. Secondly, the framework could serve as a complementary tool to the bottom-up approach for liquidity risk stress testing. This is particularly so in view of the difficulty to incorporate contagious default risk under the bottom-up approach.4 By contrast, default

32、 risk of banks is endogenised in this framework and contagious default risk is thus possible through interbank and capital markets. The proposed framework can be readily applied to other banking system as the required input data are publicly available. The reminder of the paper is organised as follows. The stress-testing framework is outlined in the following section. Sections