what network science has to say about large universal banks

A few days after he made the fateful acquisition of Golden West, Ken Thompson, then CEO of Wachovia Corporation exuberantly celebrated the universal bank model that had emerged in the United States in the wake of the repeal of the 1933 Glass-Steagall Act:

[T]here is great value in the universal bank model for both customers and shareholders. . . . [I]t is not an easy model to execute, but if you can knit disparate businesses together in a way that brings great value to customers, you will undoubtedly also deliver exceptional value to shareholders. . . . Done right, size enhances competitive power. . . . [A] universal bank . . . has much more to offer customers. . . . With economies of scale, a company can better afford the technology and longer branch hours that customers demand.

This view has been repeated by many large bank CEOs over recent years, even after the Financial Crisis. The repeal of Glass-Steagall permitted traditional and investment banks to form conglomerates, and the supposed diversification of earnings and risk that these conglomerates would seem to offer has been cited as their principal advantage.

Yet the juxtaposition of traditional and investment banking is thought by many, including some of the most experienced and senior regulators such as Paul Volcker, to be one of the core reasons for the financial meltdown. Clearly we need scientific research to help determine who is right. In my view, the most promising area of research is complexity science because financial systems, the regulatory ecology, and financial organizations themselves are themselves all highly complex. The stability of these systems, macro and micro, the contagion that can spread within them, and their tight coupling and dynamic interdependency exhibit all the phenomena that complexity scientists have spent much time studying.

I have just come across an intriguing paper published by a group of scientists out of the New England Complex Systems Institute (NECSI), called “Networks of Economic Market Interdependence and Systemic Risk.” This paper studies the dynamic network of relationships among corporations that underlie what are termed “cascading economic failures” (cascading failure being common to many complex systems). The economic risks associated with cascading financial losses are normally ignored by conventional economic analysis, which tends to take into account only linear exogenous events.

Using network analysis, the authors present research that strongly confirms what many of us have thought all along, namely that the clustering of financial services in the wake of the repeal of Glass-Steagall in 1999, and the rapid rise of universal banks in the wake of that repeal, made our financial system much more vulnerable to collapse. Before 1999, financial relationships were much more segregated, not only by the Glass-Steagall Act but also by the Bank Holding Company Act of 1956 (which imposed strict limits on the ability of financial and non-financial companies to form conglomerates, but which was watered down through regulatory interpretations in the age of deregulation). The NECSI study leads its authors to conclude that “segregating financial relationships, particularly among activities that are not otherwise related, or are weakly related, reduces systemic risk,” and to observe that

One of the arguments in favor of deregulation was that banks, by investing in diverse sectors, would have greater stability. Our analysis implies that the investment across economic sectors itself creates increased cross-linking of otherwise much more weakly coupled parts of the economy, causing dependencies that increase, rather than decrease, risk. Quite generally, separation prevents failure propagation and connections increase risks of global crises.

This study suggests that erosion of the Glass–Steagall Act, the consolidation of banking functions, and cross sector investments eliminated “firewalls” that could have prevented the housing sector decline from triggering a wider financial and economic crisis.

Although they are agnostic about the cost-benefit tradeoffs that such separation would incur, it seems to me that the onus is once again on the proponents of our heavily subsidized universal banking system to demonstrate–with science, not bluster–why the advantages of the model they support would outweigh the greatly enhanced systemic risk these conglomerates appear to generate.

So far they have yet to produce any.

3 comments for “what network science has to say about large universal banks

Leave a Reply

Your email address will not be published.