The Bankwatch

Tracking the consumer evolution of financial services

Banking is a protected industry and that protection results in limited competition and little innovation

Andrew Haldane (Andrew Haldane – Executive Director for Financial Stability and member of the Financial Policy Committee) is a favourite.  His talks are rich in history and assessment of patterns that point to future solutions. 

This talk earlier this month was presented at the Securities Industry and Financial Markets Association’s (SIFMA) Legal Entity Identifier Symposium in New York, and is both very good in the Haldane tradition but also a direct shot at the poor state of bank technology and how that contributed to the 2008 crisis.

His overall point is lack of transparency of risk within financial services.  A solution he suggests is language, not spoken but in the broader sense of communication.  This language solution exists in two areas that he explores – supply chains and the world wide web.  Both are examples of network activity as are financial services.  The comparison notes the efficiency in the first two and the opposite in financial services.  The reason is the common worldwide language in the first two and the lack of that in financial services.

He notes that lack of effective communication results in instant doubt whenever there is an emergency such as 2008, and that translates into each bank not trusting each other so interbank activity dries up and the markets freeze.  This also occurred in 2011 in Europe.

The communication that he speaks of would mean banks would instantly know in real time the status of their next in line transacting partners and their partners (counterparties and their counterparties).  As he notes in 2008 the “The whole credit chain was immersed in fog”.

News Release – Towards a common financial language – speech by Andrew Haldane | BofE

Very few firms, possibly none, had the information systems necessary to aggregate quickly information on exposures and risks. This hindered effective consolidated risk management. For some of the world’s biggest banks that proved terminal, as unforeseen risks swamped undermanned risk systems.

These problems were even more acute across firms. Many banks lacked adequate information on the risk of their counterparties, much less their counterparties’ counterparties. The whole credit chain was immersed in fog. These information failures contributed importantly to failures in, and seizures of, many of the world’s core financial markets, including the interbank money and securitisation markets.

In seeking examples to highlight what the financial system needs he points to supply chains and the www.

To gauge that, we begin by discussing the progress made by two well-known industrial networks over recent years – product supply chains and the world wide web. For both, a common language was the prime-mover behind a technological transformation. That transformation has delivered huge improvements in system resilience and productivity.

Compared with these industries, finance has been a laggard. But the tide is turning. Since the crisis, significant progress has been made by financial firms, regulators and trade associations in developing information systems and the common data standards necessary to underpin them.

The introduction of the barcode vs bespoke company specific languages in 1974 was revolutionary.

During the 1980s and 1990s, this UPC technology extended its reach along the supply chain.  It did so by developing a system for locating products as well as identifying them. A Global Location Number (GLN) was introduced in 1980, tying products to fixed locations. And in 1989, the Serial Shipping Container Code (SSCC) was devised, enabling products to be grouped in transit.

Between 1980 and 2005, the total value of world exports quintupled.6 A common language was their glue.

Then a short history of the web and the glue that binds it.

On October 29 1969, the first-ever message was sent between computer networks at UCLA
and Stanford Universities.

In March 1989, Tim Berners-Lee, a British software engineer working at CERN, the
Geneva-based nuclear research lab, wrote a memo to his boss Mike Sendall.7 It was a
suggestion for improving information management at CERN

At the centre of this web was a common language – Hyper-Text Markup Language (HTML). This was a new lingua franca for computers, allowing them to communicate irrespective of their operating systems and underlying technologies.

To accompany this new language, Berners-Lee created some new co-ordinates for this global web – Universal Resource Identifiers, the most well-known of which is the URL (Uniform Resource Locator).

Looking at the measurement of the benefits from standardised communication, there are dramatic results in supply chains.

The risk of being out-of-stock in Wal-Mart stores equipped with radio frequency technology has been found to be 30% lower than in stores without that technology

His example of Echo Bay Technology (eBay) is a little weaker and we have seen eBay shift to more of a shopping mall model, but the user feedback model is powerful.

… Omidyar adapted the site to allow buyers and sellers to give each other feedback on transactions, positive or negative. These scores created a method for buyers to assess risk.

These examples are all in contrast to banks.  They have none of that efficiency.

The product recall example is powerful and provides a powerful metaphor for banks who trade products.  in todays world when a bank or a banks product is in question, the immediate reaction has to be to shut down all trading, because it is impossible to assess the extent of the risk.  It is systemic and cannot be localised whereas the drug manufacturers in his example have solved that.

If a product needs to be recalled, the manufacturer updates the common data pool with an instruction, alerting the distributor or dispenser at the next point in the supply chain. If a recalled product has already been sold, the points of sale can be tracked and those affected can be notified. Panic is no longer systemic; it is quickly localised.

Next up is the inter country work to develop standards for Product Identifier (PI) and Legal Entity identifier (LEI) but finance has a long way to go.  Such identifiers would identify banks and bank products.

He notes the extent of manual intervention within finance which exacerbates the opportunity for error.  Manual in this context includes Microsoft Excel which remains the current standard for much of finance.  Excel is both manual and fragmented.

A fragmented data infrastructure has a number of nasty side-effects. One is a reliance on manual intervention to operate data and risk management systems. Complete automation of business processes – second nature along the product supply chain – is a rare bird in banking. As well as improving efficiency, automation reduces the possibility of misreporting of exposures through human error.

Now Haldane gets to the roots of the problem in finance that this introduction might solve.  He compares the days that it took for banks’ to assemble consolidated risk pictures versus the real time view that exists in Wal-Mart.  The specific example of Lehmans losing a real estate investment exposure of $6Bn highlights this.

Piecemeal data capture also complicates aggregation of risk exposures. That in turn undermine effective consolidated risk management. In its 2010 report, the Senior Supervisors Group found that some firms could take weeks to aggregate exposures across their balance sheet.34 Even the best-run firms took several hours or days. This, too, contrasts starkly with the real-time inventory management systems operating along manufacturing supply chains, not only among the Wal-Marts but among the corner stores.

These data and risk management problems were well-highlighted in the official report into the
failure of Lehman Brothers, a firm whose risk management was felt to be close to the frontier
of risk management best practices ahead of the crisis.35 For example, the mistaken exclusion
of a significant commercial real estate exposure of $6 billion meant there was serious
misreporting to the Lehman’s board during the crucial months prior to its failure. Information
on a significant exposure was lost in translation.

Similarly with Lehmans, while bank wide exposures were within limits individual components were within breach and this was not immediately evident.

He notes this applies to all banks and not just Lehmans.

Real-time, consolidated data capture is still a distant aspiration. As long as that is the case, real-time consolidated risk management remains a pipe dream.

Not only is the risk management in banks hindered by lack of technology solutions he notes the possible cause when he looks at barriers to entry.

Lowering barriers to entry

Paul Volcker famously commented that the only useful piece of technological innovation in banking over the past 30 years had been the ATM. What is certainly true is that some markets for banking products appear to lack contestability. That has meant large incumbents have often dominated the banking landscape, in much the same way as was true until recently in music and publishing. In banking, this lack of contestability is one root cause of the too-big-to-fail problem.

Consistent with that, birth and death rates in banking are lower than among non-financial companies. They are lower even than in other areas of finance. Death rates among US banks have averaged around 0.2% per year over the past 70 years. Among hedge funds, average annual rates of attrition have been closer to 7% per year. Birth rates are similarly low. Remarkably, up until 2010 no new major bank had been set up in the UK for a century.

Banking is a protected industry and that protection results in limited competition and little innovation.  He makes the point that a lack of common language hinders new entrants.  We see that today with P2P lending platforms.  He sugests barriers to entry need to be lowered so that companies such as Paypal, and Zopa can be allowed to grow and develop better models than existing banks.

Commercial peer-to-peer lending, using the web as a conduit, is an emerging business. For example, in the UK companies such as Zopa, Funding Circle and Crowdcube are developing this model. At present, these companies are tiny. But so, a decade and a half ago, was Google. If eBay can solve the lemons problem in the second-hand sales market, it can be done in the market for loans.

With open access to borrower information, held centrally and virtually, there is no reason why end-savers and end-investors cannot connect directly. The banking middle men may in time become the surplus links in the chain. Where music and publishing have led, finance could follow. An information web, linked by a common language, makes that disintermediated model of finance a more realistic possibility.

Relevance to Bankwatch:

It’s a fascinating talk.  Basically Haldane suggests:

  1. A root problem of banks is a lack of consistent entity assessment and of product assessment
  2. This root problem means there is no ready means to assess the risk of other entities (banks, funds etc) or the risk of the products those entities produce
  3. A second root problem lies in banks lack of technology innovation.  Banks are as fragmented internally, as they are fragmented amongst each other. This fragmentation is especially evident in their technology.
  4. This second root problem means that even with the consistent entity and product assessment problem solved, it would be impossible to effectively share that with each other and with the regulators.

Finally he notes that perhaps the best solution is to encourage new entrants who (I would suggest) are more skilled in technology, and will be better placed to solve these problems from the outside and in so doing perhaps force the incumbents to change.

A basic example I would suggest is wire payments.  Apart from the anachronistic name, the SWIFT system employed by banks, along with their own internal processes, results in money taking up to 48 hours to reach a destination bank.  This in contrast to the problem Wal-Mart have in moving physical inventory.  Surely the matter of sending money instantly would be a simple problem to solve but not without trusted standards.

In some ways this harks back to this which I wrote in 2008 suggesting there will be two kinds of banks.  Innovators and utilities.  I stick with that.  In this view most of todays banks would become utilities providing basic banking functionality that keeps the money moving.  High risk investment vehicles will be ring fenced outside these utilities. 

Innovators will be able to enter with transparent models which are easily understood due to real time data that does a significantly better job of providing risk profiles that ensure the customers and the regulators are comfortable with the innovation.

Written by Colin Henderson

March 24, 2012 at 21:45

Posted in Uncategorized

One Response

Subscribe to comments with RSS.

  1. […] banking directCredit, eh. Canada's Credit Card & Personal Finance BlogThe Bankwatch […]


Comments are closed.

%d bloggers like this: