Determinants of Cross-border Banking System Exposures

Third, the impact of interest rate spreads on the risk transfers requires further investigation. Cerutti, 2014 reports that TED spread (the spread between three-month US dollar Libor and the three-month US treasury bill rate) has negative effect on the aggregated foreign credit exposure of the internationally active banks. Our model shows that at the individual level the same effect of interest rate spread can be observed. However, the reason for such a link remains unclear.

Finally, our model confirms the hypothesis that cross-border exposures are indeed statistically linked to a number of macroeconomic indicators of creditor and borrower countries. Inflation, debt to gdp ratio and financial leverage are all statistically significant with an expected sign associated with them.

The main policy recommendation that can be made on the basis of our findings is for countries to follow the path of upholding macroeconomic stability to make sure that the excess funding from other countries flows towards them.

Conclusion

Global financial system with increased role of cross-border transactions requires new analytical tools for policymakers concerned with financial stability. Our paper has provided an example of two powerful statistical tools, which can be used to better understand the way the capital flows in and out of the country.

Network analysis and its ability to grasp tight connections between the financial systems of different countries can produce meaningful insights about the possibilities of contagion and risks to financial stability. Our analysis showed that since the great recession and deleveraging that followed, some of the indicators start increasing again and while the levels seen right before the crisis have not been tested again, it requires an investigation by the financial authorities. Furthermore, we show that the density of the financial connections has stayed at the same levels in the entire time period under consideration. It suggests that the same contagion channels that destabilized the world economy in 2008-2009 are still in place and many precautions, including stress testing and macroprudential regulation should be reinforced to make sure that the same damage cannot be done in the future. Finally, our research highlights certain countries and connections that have an outsized influence at the system as a whole. They should be considered systemically important and require additional attention.

Panel data modelling, on the other hand, points to the risk factors that are most likely to cause a shock to the system. These factors are inflation, debt to GDP ratio and financial leverage of the banking sector of the borrower country. A rapid change for the worse in any of these indicators could have repercussions across the system. We also find that the driving forces behind the risk transfers in form of derivatives differ from the ones impacting the total exposures. We argue that future research should include different model specifications for these two types of exposures. Another suggestion for future research is an exploration of a link between interest rate spread and cross-border risk transfer. Cerutti, 2014 and our research suggest that the sign in front of interest rate spread is positive both at the aggregate and individual level, respectively.

To conclude, it is important to reiterate just how powerful the combination of network analysis and panel data modelling can be in the financial stability research. Networks can suggest possible contagion channels and panel regressions point towards the most significant risk factors, which means that the approach to financial stability policy can be finetuned for each financial system in the network. There are currently multiple projects dedicated to the same idea, most noticeably “FINEXUS Leverage Network Framework for Stress-testing” project at Center for Financial Networks and Sustainability at the University of Zurich. Hopefully, these efforts by the research community will result in a more informed and effective financial policy.

This work can be continued in the following directions. First, if more data becomes available about the risk transfers to and from the periphery countries, their data can also be included in the analysis. Second, further research into the driving forces behind the derivatives exposures can be conducted. Finally, a more complexed model than the panel regression can be used to model the changes in exposures, e.g. global vector autoregression model.

References

1.Minoiu C., Reyes J.A. A network analysis of global banking: 1978-2010 // Journal of Financial Stability. 2013. Vol. 9. № 2. P. 168-184.

2.Baier S.L., Bergstrand J.H., Feng M. Economic integration agreements and the margins of international trade // Journal of International Economics. 2014. Vol. 93. № 2. P. 339-350.

3.BIS. Reporting guidelines for the BIS international banking statistics. , 2019.

4.BIS (2020), Effective exchange rate indices. URL: https://www.bis.org/statistics/eer.htm?m=6%7C381%7C676 (Accessed on 25 April 2020).

5.Bonacich P. Power and centrality: A family of measures //American journal of sociology. - 1987. - Vol. 92. - №. 5. - P. 1170-1182.

6.Brandes U. A faster algorithm for betweenness centrality //Journal of mathematical sociology. - 2001. - Vol. 25. - №. 2. - P. 163-177.

7.Cerutti E. Drivers of cross-border banking exposures during the crisis // Journal of Banking and Finance. 2014. Vol. 55. P. 340-357.

8.Cerutti E. Drivers of cross-border banking exposures during the crisis // Journal of Banking and Finance. 2014. Vol. 55. P. 340-357.

9.Cerutti E. Drivers of cross-border banking exposures during the crisis // Journal of Banking and Finance. 2014. Vol. 55. P. 340-357.

10.Cerutti E., Claessens S., McGuire P.M. Systemic Risk in Global Banking: What Can Available Data Tell Us and What More Data are Needed? // SSRN Electronic Journal. 2012. № March 2019.

11.Croissant Y, Millo G (2018). Panel Data Econometrics with R: the plm package. Wiley.

12.Csardi G, Nepusz T: The igraph software package for complex network research, InterJournal, Complex Systems 1695. 2006. http://igraph.org

13.Csardi G, Nepusz T: The igraph software package for complex network research, InterJournal, Complex Systems 1695. 2006. http://igraph.org

14.Engel J., Pagano A., Scherer M. A block-structured model for banking networks across multiple countries. , 2019b.

15.Engel J., Pagano A., Scherer M. Reconstructing the topology of financial networks from degree distributions and reciprocity // Journal of Multivariate Analysis. 2019a. Vol. 172. P. 210-222.

16.Finding a locally optimal solution in hopes of stumbling upon a globally optimal one.

17.Freeman L. C., Borgatti S. P., White D. R. Centrality in valued graphs: A measure of betweenness based on network flow //Social networks. - 1991. - Vol. 13. - №. 2. - P. 141-154.

18.Garlaschelli D., Loffredo M.I. Generalized Bose-fermi statistics and structural correlations in weighted networks // Physical Review Letters. 2009. Vol. 102. № 3. P. 2-5.

Appendix

A.1. Pair regressions model summary (total exposure data).

A.2. Pair regressions model summary (derivatives data).

A.3. Panel regressions model summary

A.4. Strength (in) of individual countries (total exposure data)

A.5. Strength (out) of individual countries (total exposure data)

A.6. Average neighbor strength of individual countries (total exposure data)

A.7. Centrality measures of individual countries (total exposure data)

A.8. Cross-border risk transfers network for 2019Q4 (total exposure data)

A.9. Cross-border risk transfers derivatives network for 2019Q4

local economy international

Размещено на Allbest.ru