(Corrects first paragraph to reflect that Google doesn’t need to act.)
Aug. 7 (Bloomberg) -- The mathematical insight that turned Google Inc. into a multibillion-dollar company has the potential to help the world avert the next financial crisis. If only banks made public the data required to do the job.
Sixteen years ago, the founders of Google -- computer scientists Larry Page and Sergey Brin -- introduced an algorithm to measure the “importance” of Web pages relative to any set of keywords.
Known as PageRank, it works on the notion that Web pages effectively vote for other pages by linking to them. The most important ones, Page and Brin reasoned, should be those drawing links from many other pages, especially from other really important ones.
If this definition sounds circular, it is. It also captures an authentic reality, which is why respecting it gives far superior results. Page and Brin’s breakthrough involved using mathematics to make it work. The required ideas don’t go much beyond high-school algebra, although it takes lots of computing power to make something as sprawling as the World Wide Web possible.
What could this have to do with finance? Quite a lot. The systemic risk that turned the U.S. subprime-lending crisis into a global disaster is circular, too. We can’t identify it simply by looking for the banks with the most assets or the biggest portfolios of risky loans. What matters is how many links a bank has to other institutions, how strong those links are and how risky those other banks are, not least because they too have links to other risky banks.
Something like PageRank might be just the right thing to cut through it. That’s the argument, at least, made by a team of European physicists and economists in a new study. Their algorithm, DebtRank, seeks to measure the total economic value that would be destroyed if a bank became distressed or went into default. It does so by moving outward from the bank through the web of links in the financial system to estimate all the various consequences likely to accrue from one failure. Banks connected to more banks with high DebtRank scores would, naturally, have higher DebtRank scores themselves. (I have put a little of the technical detail on my blog.)
As a demonstration, the researchers calculated DebtRank on the basis of the known network of equity investments linking institutions -- pretty much the best they can do with publicly available data. If Bank A owns stock in Bank B, the two are linked. This network, of course, reflects only a subset of the many links created by derivatives and other instruments, so the calculation is a little like working out the best driving route from New York to Los Angeles while ignoring two-thirds of all the roads. Nevertheless, it’s useful for demonstrating what might be possible with more complete data.
The analysis offers some surprises. At the peak of the financial crisis, in November 2008, for example, DebtRank scores for the largest 20 or so banks show that simple bank size isn’t as important as we have come to think. Institutions such as Barclays Plc, Bank of America Corp., JPMorgan Chase & Co. and Royal Bank of Scotland Group Plc presented more systemic risk than did Citigroup Inc. or Deutsche Bank AG, despite being significantly smaller in total assets. Wells Fargo & Co. stands out even more: It presented as much systemic risk as Citigroup, despite having only a quarter of the assets.
An algorithm alone can’t save the world, and this isn’t the final word on the best way to measure systemic risk. Yet the apparent superiority of the DebtRank approach underscores how our ability to monitor the financial system depends wholly on the availability of data. Currently, most of the information that would be needed to calculate DebtRank or any other similar measure is simply not public.
Imagine a world in which banks and other financial institutions were legally required to disclose absolutely all of their assets and liabilities to central banks, which would in turn make that information public on a website. Regulators -- indeed, anyone -- would then be able to see the whole network and assess a bank’s situation in full clarity. Anyone so inclined could calculate measures such as DebtRank and assess how much any particular bank is contributing to potential financial instability.
With full transparency, it’s just possible that the core business of lenders would go back to assessing the creditworthiness of borrowers. They would need to do so to maintain a good reputation and to borrow themselves, as any risky loans they made would be known to all. In such a situation, the economist and physicist Stefan Thurner of Medical University of Vienna suggests, “financial institutions would only survive and prosper if they assess the risk of others better than their peers.”
That is a radical idea, so radical it is almost certainly a political nonstarter. But as the British physicist William Thomson, also known as Lord Kelvin, put it back in the 19th century: “What you cannot measure, you cannot hope to improve.” It’s a lasting piece of wisdom.
(Mark Buchanan, a theoretical physicist and the author of “The Social Atom: Why the Rich Get Richer, Cheaters Get Caught and Your Neighbor Usually Looks Like You,” is a Bloomberg View columnist. The opinions expressed are his own.)
Read more opinion online from Bloomberg View. Subscribe to receive a daily e-mail highlighting new View editorials, columns and op-ed articles.
Today’s highlights: the editors on how governments should brace for rising sea levels and on Knight Capital and computerized trading; Jeffrey Goldberg on anti-Semitism in Egypt; William Pesek on rebuilding the tsunami zone with microfinance; Ramesh Ponnuru on improving political TV shows; Betsey Stevenson and Justin Wolfers on the booming business in empirical economics; Richard Cohen on whether fourth is good enough at the Olympics; Caleb Scharf on how massive black holes regulate star creation.
To contact the writer of this article: Mark Buchanan at firstname.lastname@example.org.
To contact the editor responsible for this article: Mark Whitehouse at email@example.com.