In this paper we explore the developments in the cross-border bank exposures using the BIS International Banking Statistics. To this end, we treat the web of the cross-border bank exposures as a "network" and investigate the characteristics of the network topology, and compute various statistical measures for the network topology. We find that the network of cross-border bank exposure has become more tightly connected over time. The network now has higher connectivity, a shorter average path length, a higher average degree, and a higher clustering coefficient than in the past. In particular, we observe that such tendency has never been hampered by any disturbances or crises in international financial markets (such as the East Asia currency crisis in 1997 or the LTCM near-default event in 1998). We see both costs and benefits from these developments in the cross-border bank exposures. On the one hand, systemic risk in international financial markets is likely to increase because of the more direct and more widely spreading spillover effects of a crisis in one country once it occurs. On the other hand, efficiency of international financial markets is expected to further improve in terms of capital and risk allocation.
[1]
Goetz von Peter,et al.
International Banking Centres: A Network Perspective
,
2007
.
[2]
Bethany S. Dohleman.
Exploratory social network analysis with Pajek
,
2006
.
[3]
Matthew O. Jackson,et al.
The Economics of Social Networks
,
2005
.
[4]
L. Summers.
International Financial Crises: Causes, Prevention, and Cures
,
2000
.
[5]
W. C. Hunter,et al.
Global financial crises : lessons from recent events
,
2000
.
[6]
Vladimir Batagelj,et al.
Pajek Program for Analysis and Visualization of Large Networks
,
2007
.
[7]
Hideki Takayasu,et al.
Fractal Network derived from banking transaction -- An analysis of network structures formed by financial institutions --
,
2004
.
[8]
Walter E. Beyeler,et al.
The topology of interbank payment flows
,
2007
.