CONCLUSION AND FURTHER WORK

The main conclusions we can extract from this work is that although in recent times the topic has received a well-deserved importance into the regulatory arena, there are relevant aspects which we believe should not be ignored.

For some time the literature adhered to the be­lief that the topology of the network was enough to characterize the systemic riskiness of a particular financial system. Previous work (Marquez-Diez- Canedo, et al., 2009) shed some light about that issue and pointed at the fact that is not only the topology but also the size and concentration of the exposure.

This exercise brings to light one more issue, the relevance of the initial macroeconomic shock. It also carries one more argument against the idiosyncratic, one-by-one, evaluation of failures: it is impossible to dissociate banks from their economic environment. Pretending a “caeteris paribus” failure would be misleading.

Another important conclusion is that, to some extent, it could be the case that size does not matter. The focus has been traditionally on larger banks, and rightly so, but it should not be at the cost of disregarding smaller banks. As evidenced by the case study, a severe enough shock can weaken the whole system to the extent that just a little push can drop the first dominoes.

There are still issues that need to be addressed. The first one is, these simulation exercise may be understating the systemic effect. After all, this is the one-month effect of a shock, especially mis­leading in credit, where losses usually take longer to take place. The issue here is how to deal with the implicit trade-off: a more accurate credit risk measurement requires a sacrifice in the market risk measurement accuracy and vice versa.

Another issue is the weight of the tails of the parametric distribution. To keep methodological consistency, the underlying assumption was that of normality in the shocks even though we know that extreme events tend to occur more often than predicted by the light-tailed normal distribution.

Finally, one needs to acknowledge a limitation of this framework in terms of policy analysis: cur­rent infrastructure allows obtaining quite accurate market and credit risk measurement given the financial position of the participants of the system. However, a more complete policy analysis requires the ability to analyze “what if” cases, which neces­sary need to consider that financial positions are not exogenous, that are an endogenous element that can also impact the risk factors. Moreover, bankruptcy cost are not considered in the current framework but it is important as bankruptcy cost could deplete even more bank’s assets causing larger losses to other banks and this in turn could alleviate the necessity of very large initial shocks in order for triggering contagion.

ACKNOWLEDGMENT

The authors would like to express their gratitude to Pascual O’Doherty Madrazo and Juan Pablo Graf Noriega for supporting this research. The authors are also grateful to Juan Pablo Solorzano Margain, Gent Bajraj, Pablo Cesar Vazquez Espi­nal, Fernando Bizuet Cabrera and Eduardo Ponce Alonso for their research assistance. Finally, this work has benefited from the comments of anony­mous referees, Andrew J. Patton and the audience at the BIS CCA Conference on “Systemic risk, bank behaviour and regulation over the business cycle” Buenos Aires, 18-19 March 2010.

REFERENCES

Adrian, T., & Brunnermeier, M. K. (2008). Covar staff reports 348. New York, NY: Federal Reserve Bank of New York.

Aikman, D., Alessandri, P., Eklund, B., Gai, P., Kapadia, S., & Martin, E.... Willison, M. (2009). Funding liquidity risk in a quantitative model of systemic stability. Working paper 372. London, UK: Bank of England.

Allen, F., & Gale, D. (2000). Financial contagion. The Journal of Political Economy, 108, 1-33. doi:10.1086/262109

Barnhill, T. M., & Souto, M.

Bartram, S. M., Brown, G. W., & Hund, J. E. (2007). Estimating systemic risk in the inter­national financial system. Journal of Financial Economics, 86(3), 835-869. doi:10.1016/j. jfineco.2006.10.001

Battiston, S., Gatti, D. D., Gallegati, M., Green­wald, B., & Stiglitz, J. E. (2009). Liaisons dan- gereuses: Increasing connectivity, risk sharing and systemic risk. Paper presented at Networks, Power and Relations: First International NPR Workshop Catholic University of the Sacred Heart. Milan, Italy.

Blavarg, M., & Nimander, B. (2002). Interbank exposures and systemic risk. Sveriges Riskbank Economic Review, 2, 19-45.

Borio, C., Drehmann, M., & Tsatsaronis, K. (2012). Stress-testing macro stress testing: Does it live up to expectations? BIS Working Paper 369. Basel, Switzerland: Bank for International Settlements.

Boss, M., Elsinger, H., Summer, M., & Thurner,

S. (2004a). An empirical analysis of the network structure of the Austrian interbank market. In Oesterreichische Nationalbank Financial Stabil­ity Report (pp. 77-87). Vienna, Austria: Oester- reichische Nationalbank.

Boss, M., Elsinger, H., Summer, M., & Thurner,

S. (2004b). The network topology of the inter­bank market. Quantitative Finance, 4, 677-684. doi:10.1080/14697680400020325

Boss, M., Krenn, G., Puhr, C., & Summer, M. (2006). Systemic risk monitor: A model for sys­temic risk analysis and stress testing of banking systems. Financial Stability Report, 11, 83-95.

Cihak, M. (2007). Introduction to applied stress testing. IMF Working Paper WP/07/59. Washing­ton, DC: International Monetary Fund.

Cocco, J. F., Gomes, F. J., & Martins, N. C. (2009). Lending relationships in the interbank market.

Degryse, & Nguyen. (2004). Interbank exposures: An empirical examination of systemic risk in the Belgian banking system. Discussion Paper 2004­04. Tilburg, The Netherlands: Tilburg University.

Enders, W. (2004). Applied econometric time series. New York, NY: Wiley.

FSB. (2009). Guidance to assess the systemic importance of financial institutions, markets and instruments: Initial considerations. Report. Washington, DC: Financial Stability Board.

Furfine, C. H. (1999). Interbank exposures: Quantifying the risk of contagion. BIS Working Papers 70. Basel, Switzerland: Bank for Interna­tional Settlements.

Goodhart, C. A. E., Sunirand, P., & Tsomocos, D. P. (2006). A model to analyse financial fragility. Economic Theory, 27, 107-142. doi:10.1007/ s00199-004-0572-7

Graf, J. P., Guerrero, S., & Lopez-Gallo, F. (2005). Interbank exposures and contagion: An empirical analysis for the Mexican banking sector. Mexico City, Mexico: Banco de Mexico.

Hamilton, J. D. (1994). Time series analysis. Princeton, NJ: Princeton University Press.

Iori, G., de Masi, G., Precup, O. V., Gabbi, G., & Caldarelli, G. (2005). A network analysis of the Italian overnight money market. Discussion Paper Series 05/05. London, UK: City University London.

Lehar, A. (2005). Measuring systemic risk: A risk management approach. Journal of Banking & Finance, 29(10), 2577-2603. doi:10.1016/j. jbankfin.2004.09.007

Markose, S. M., Giansante, S., Gatkowski, M., & Shaghagi, A. R. (2009). Too interconnected to fail: Financial contagion and systemic risk in network model of CDS and other credit enhance­ment obligations of US banks. Paper presented at the ECB Workshop Recent Advances in Modeling Systemic Risk Using Network Analysis. Frankfurt, Germany.

Marquez-Diez-Canedo, J., & Martfnez-Jaramillo,

S. (2007). A network model of systemic risk: Stress testing thefinancial system.

Marquez-Diez-Canedo, J., Martfnez-Jaramillo, S., & Perez-Perez, O. (2009). A network model of systemic risk: Stress testing the banking system. Journal of Intelligent Systems in Accounting, Finance & Management, 16, 87-110.

Martfnez-Jaramillo, S., Perez, O., Avila, F., & Lopez-Gallo, F. (2010). Systemic risk, financial contagion and financial fragility. Journal of Economic Dynamics & Control, 34, 2358-2374. doi:10.1016/j.jedc.2010.06.004

Mistrulli, P. E. (2007). Assessing financial contagion in the interbank market: Maximum entropy versus observed interbank lending pat­terns. Economic working papers 641. Rome, Italy: Bank of Italy.

Muller, J. (2006). Interbank credit lines as a chan­nel of contagion. Journal of Financial Services Research, 29, 37-60. doi:10.1007/s10693-005- 5107-2

Nier, E., Yang, J., Yorulmazer, T., & Alentorn, A. (2006). Network models and financial stabil­ity. Journal of Economic Dynamics & Control, 31, 2033-2060. doi:10.1016/j.jedc.2007.01.014

Rochet, J.-C. (2009). An industrial organization approach to the too big to fail problem. Paper presented at the Federal Reserve Bank of Boston 54th Economic Conference. Boston, MA.

Segoviano, M. A., & Goodhart, C. (2009). Banking stability measures. IMF Working Pa­pers WP/09/04. Washington, DC: International Monetary Fund.

Sheldon, G., & Maurer, M. (1998). Interbank lending and systemic risk: An empirical analysis of Switzerland. Swiss Journal of Economics and Statistics, 134, 685-704.

Sorge, M. (2004). Stress-testing financial sys­tems: An overview of current methodologies. BIS Working Paper 165. Basel, Switzerland: Bank for International Settlements.

Sorge, M., & Virolainen, K. (2006). A com­parative analysis of macro stress-testing meth­odologies with application to Finland. Journal of Financial Stability, 2, 113-151. doi:10.1016/j. jfs.2005.07.002

Tarashev, N., Borio, C., & Tsatsaronis, K. (2009, September).

Toivanen, M. (2009). Financial interlinkages and risk of contagion in the Finnish interbank market. Discussion Papers 6/2009. Helsinki, Finland: Bank of Finland Research.

Upper, C. (2007). Using counterfactual simula­tions to assess the danger of contagion in inter­bank markets. BIS Working Papers 234. Basel, Switzerland: BIS.

Upper, C., & Worms, A. (2004). Estimating bilateral exposures in the German interbank market: Is there a danger of contagion? European Economic Review, 48, 827-849. doi:10.1016/j. euroecorev.2003.12.009

Wells, S. (2002). UK interbank exposures: Systemic risk implications. Financial Stability Review, 175-182.

10

11

ENDNOTES

The opinions expressed here are those of the authors and do not represent the views of the Mexican Central Bank.

This definition is shared by Marquez-Diez- Canedo et al. (2009) and Rochet (2009).

Particularly in banking systems through the interbank market.

See Upper (2007) for a good survey on the literature.

Indirect contagion usually refers to similar business models or implicit correlations. Overexposure is explained in detail in Martinez-Jaramillo et al. (2010) and refers to banks that are exposed on the interbank exposures network by an amount, which could bring their capital ratio below the regulatory minimum.

See the note in the Financial Times “B ankers to Lobby for Softer Reforms” on January 24th 2010.

In the last few years, the numbers of banks in Mexico has been increased.

In fact, we have data from previous years, but due to differences on the quality of such information, we can only fully trust the data from January 2005.

See Mistrulli (2007) for a comparison of contagion in a network built under the maximum entropy principle and without such assumption.

The analysis of extreme conditions or “stress scenarios’ ’ can be tricky using VAR models, especially when underlying the model there is a normal distribution of shocks. The nor­mality assumption implies that extreme, or sometimes not even so extreme, realizations occur with probability virtually zero. The problem with this situation is that experience tells that shocks tend to have heavier tails

12

13

14

15

than those predicted by normal distributions and these “extreme” virtually impossible realizations occur much more often than the

model predicts; hence, it is highly relevant to assess, regardless of the projected probability of occurrence, the effect of these scenarios. Notice that the white noise assumption considerably simplifies this exercise.

In Marquez-Diez-Canedo et al. (2009) the number of randomly generated scenarios was five million.

In a strict sense, the probability of such extreme events is in a different order of magnitude.

The CoVaR should not be confused with the

17

18

19

20

CVaR (Conditional VaR).

Alternatively, the CoVaR can be derived by quantile regression, as done in Adrian and Brunnermeier (2008).

21

22

23

24

25

In a slight abuse of terminology, we will refer to this kind of event as a “failure.” For instance, within the context of stress testing, VAR models reportedly perform poorly near the tails of the distribution.

In our case, the most recent available infor­mation.

The market risk-measuring infrastructure used by Banco de Mexico was used to per­form the valuation of the portfolio at each different scenario.

All expressed in differences in logs. Experience in the Central Bank indicates that this variable can be important for modeling during stress scenarios.

As usual, one of these dummies was dropped. See Hamilton (1994).

See Enders (2004).

This work was previously published in Simulation in Computational Finance and Economics, edited by Biliana Alexandrova- Kabadjova, Serafin Martinez-Jaramillo, Alma Lilia Garcia-Almanza, and Edward Tsang, pages 181-210, copyright 2013 by Business Science Reference (an imprint of IGI Global).