Category: Digital Disruption

System Alert – Does Not Comply With Responsible Lending!

The Royal Commission looking at Financial Services Misconduct heard today that the Commonwealth Bank’s automated system for approving overdrafts failed and so for four years from 2011 it gave some customers a line of credit they shouldn’t have received.

As a result, the volume of overdrafts rose significantly from 228,000 in 2012 (up 80% from the previous year) to 550,000 in 2014. The bank said its automated system “spat out” wrong approvals and was “doomed to fail” because of bad design.  We discuss this in our latest video blog.

In fact, questions were raised by consumer advocacy groups before the bank released there was an issue. The implementation of serviceability assessments was not made correctly. As a result of changes made to the system, the bank failed its responsible lending obligations.

It was also slow to interact with the regulator on this issue.  Once again, cultural and behavioural issues were in the spotlight.

The object lesson here is that automated credit decision systems can lead you up the garden path.  This is important given the current rush to digital channels and more automation.

Could AI Solve The Broker Problem?

Given the tenor of the Royal Commission responsible lending inquiries this week, which focussed on the complexities of brokers and lenders complying with their responsible lending obligations, we believe the future will be distinctly digital. Our banking innovation life cycle road map calls this out.

To illustrate the point, there was a timely announcement from the Opica Group who have a new, and they claim Australia’s first responsible lending engine” (RELIE). This from The Adviser.

A new artificial intelligence-based expenses verification engine has been launched for brokers and lenders to ensure responsible lending and compliance obligations are met.

Billing the tech as “Australia’s first responsible lending engine” (RELIE), the Opica Group has launched the platform to help “protect any broker or lender from a breach of their responsible lending requirements”.

According to Opica Group founder Brett Spencer, the platform is needed because “lenders traditionally have been very quick to put blame on brokers for any application that goes sour”.

Mr Spencer said that following a tighter regulatory environment and “greater scrutiny being placed on our industry by regulators”, the group identified that “brokers needed something that provided them some protection”.

As such, it built the RELIE platform to enable brokers (and lenders) to perform a “RelieCheck” that could prove they had done the adequate checks into expenses and the consumer’s ability to service the loan.

How it works

The RELIE engine makes use of a specially built artificial intelligence engine, Sherlock™, which analyses a consumer’s banking and credit card transaction data over a period of 12 months and automatically provides “income verification, an understanding of the client’s mandatory expenditure, and therefore their ability to service a loan”.

According to the group, the key differentiator of the RELIE platform when compared to credit checks is that it uses machine learning to categorise transactions, allowing for the differentiation of transaction types, including mandatory versus discretionary expenditure and recurring versus one-off spending.

It also automatically highlights areas of concerns within the transaction data such as undisclosed debts, spikes in expenditure of high-risk categories such as gambling, and possible changes in life circumstances such childbirth.

Mr Spencer commented: “With the advancements in technology and legislation crackdown, we saw an opportunity to protect brokers and automate significant components of an applicant’s income and expense verification process…

“We believe that running a RelieCheck will protect any broker or lender from a breach of their responsible lending requirements.”

Speaking to The Adviser, Mr Spencer elaborated: “While a credit check simply looks at your credit worthiness, a RelieCheck looks at the consumer’s 365-day spending and income transactions and interrogates the data from a responsible lending perspective.

“It then presents back to the broker or lender a summary of exactly what, when and where an applicant’s income and expense are positioned.”

However, the Opica Group founder said that while the AI engine “does all the grunt work” to auto categorise and allocate spends to a range of buckets (such as mandatory versus discretionary expenses), the broker is able to review each category of spend and re-allocate expenses to a different category as part of their responsible lending discussions with the customers.

Each change made is then notated by the broker in order to meet their responsible lending requirements.

Revealing that the engine has been 16 months in the making, Mr Spencer said that the group wanted to “create a platform that a broker could use to protect themselves from any unintended breach of their responsible lending requirements”.

He added: “We also wanted to speed up the physically demanding process of paper-based statement reviews so that a broker could reduce the amount of time it takes to process a loan, and in the process providing a far greater service to the customer.”

Opica Group revealed that “early indications” have shown that by performing a RelieCheck on an applicant, a broker or lender could reduce processing times by approximately 90 minutes per application (when compared to manual assessment of the applicant’s banking and credit card transactions).

Mr Spencer concluded: “We want to create a new industry standard.

“Data is a commodity, but what you do with the data is the key ingredient.”

He added that he did not believe anyone else was thinking about “what we do with the data to aid the lending process”.

Opica Group is reportedly working with a number of aggregators and lenders to establish whether the engine could be integrated into their customer relationship management (CRM) systems. The service costs $15 (plus GST) per applicant for a broker account, or $10 (plus GST) per applicant for an aggregator or lender account.

NAB launches super virtual assistant

From Financial Standards.

NAB has launched a digital assistant that helps MLC members engage with their superannuation.

Available on Google Home devices, Talk to MLC answers 15 common questions members ask: such as how to open an MLC account, find lost super and change investment options.

MLC customer experience specialist Peter Forster said the super fund expects most members to access superannuation in a way that’s convenient and personalised without the need for passwords.

He said millennials and older Australians will likely be the first to embrace Talk to MLC.

“The technology took us six weeks to develop and deploy, and we’re in the process of developing other technology at a similar speed that will help to reduce asymmetry of information and further benefit our customers,” Forster said.

He added in the near future MLC will be able to provide personalised tips to help members boost their super; project where their super balance will be at retirement time; and advise how best to invest their money in super.

NAB executive general manager of digital and innovation Jonathan Davey said the proliferation of voice-activated, hands-free devices such as Siri and Google Home and Amazon’s Alexa in the Australian market is reshaping consumer behaviour and expectations.

“We live in a world that wants instant gratification. We want quick answers and problems that are solved immediately – we don’t want to be left waiting. Our lives are busier than ever before,” Davey said.

Early this year, CBA launched Ceba, a chatbot that recognises about 60,000 consumer banking questions.

Ceba’s point of difference, according to CBA executive general manager digital Pete Steel, is that it can actually carry out tasks for customers, rather than providing instructions on how they can be done.

ANZ is also deploying chatbots with the help of Progress’ NativeChat, to enable customers to converse and transact with chatbots naturally. NASDAQ-listed Progress helps develop industry-specific and self-learning chatbots for organisations.

Banking Is Changing – A Case In Point – NAB and The Riverina

A release from NAB today.  Bye-bye branches.

In 2018, the way customers are banking in the Riverina and the surrounding areas has changed. Today, in response, NAB confirms changes to some of its branches in the area.

  • NAB invests $1.6M to improve branches in the Riverina and surrounding areas in 2017 and 2018.
  • Following consultation with local teams, NAB can confirm Ardlethan, Lockhart, Grenfell, Culcairn, Boort, Barham and Euroa branches will close in June.
  • Customers in these towns can continue to do their banking at Australia Post offices, including making deposits up to $10,000 cash or withdrawals up to $2,000 per day.
  • NAB continues to back the Riverina through its other NAB branches across the region, sponsorships, including NAB AFL Auskick, and by funding and advocating for infrastructure so regional areas can grow.
  • Our business and agri bankers will continue to service the areas.

Locally, NAB is investing more than $1.6M into improving branches in Cowra, Seymour and Kerang, completed last year, and Tatura, Alexandra and Griffith, scheduled to be completed by September 2018, including installing and upgrading 32 ATMs in the area. Many of these ATMs are ‘Smart ATMs’, where customers can make deposits, check balances, and withdraw cash so customers can bank at their convenience.

As improvements are made to some branches, other branches in the area will be closing. Between 80-90% of NAB customers in Ardlethan, Lockhart, Grenfell and Culcairn are using other branches in the area such as Temora, Wagga Wagga, Young and Holbrook. Similarly approximately 85% of customers using Euroa, Boort and Barham are using other branches .

NAB General Manager, Retail, Paul Juergens, explained the decision was a difficult one to make and was only made after careful consideration.

“While our branches continue to be an important part of what we do at NAB, the way our customers are banking has changed dramatically in recent years,” Mr Juergens said.

“Increasingly we find that our customers are banking at other branches, or prefer to do their banking online, on the phone, or through our mobile app.

“In the locations we are closing, more than 80% of our customers are also using our other NAB branches in the area.

“Importantly, we are continuing to support the Riverina and surrounding areas, including a $1.6M investment into other branches in the area as well as through local sponsorships.”

Mr Juergens emphasised that NAB wants to continue to help our customers with their banking.

“Over the coming weeks, we’ll be spending time with our customers explaining the different banking options available to them, including online banking and banking through Australia Post.

“We know that some NAB customers still like to bank in person, which is why we have a strong relationship with Australia Post offices, which offer banking services on NAB’s behalf.

“At Australia Post, NAB customers can do banking like check account balances, pay bills and make deposits up to $10,000 cash or withdrawals up to $2,000 per day.”

NAB is working with our local branch employees to discuss their next steps.

“When we make changes to our branches, we make every effort to find opportunities for our local teams at other branches in our network, and often this is possible. If we can’t find opportunities, we help our employees through The Bridge, our industry leading program where employees are provided up to six months of career coaching as they decide what’s next for them – whether that be retirement, pursuing a new career or starting a small business.”

SME Funding an Issue Says New Report

The latest edition of the Scottish Pacific SME Growth Index has been released. It gives an interesting snapshot on the critically important SME sector in Australia. Once again, as in our own SME surveys, cash-flow is king. 90% of SME owners said they faced cash-flow related issues.  That said, the non-bank sector, including Fintechs need to do more to raise awareness of the solutions they offer.

SME business confidence is on the rise finds small business owners forecasting revenue to improve during the first half of 2018.

There appears to be a splitting of the pack in SME fortunes, with a greater number of previously “unchanged” growth SMEs moving into positive or negative growth.

For most SMEs cash flow has improved compared to 12 months ago, however one in 10 say they are worse off now. The number of SMEs reporting significantly better cash flow (27%) and better cash flow (42%) will hopefully act as a major driver of new capital expenditure and business investment demand.

Despite this reported rise in cash flow, nine out of 10 SMEs say they had cash flow issues in 2017 and nine out of 10 say these issues impacted on revenue. On average, small businesses say that better cash flow would have increased their 2017 revenue by 5-10%.

For SMEs with plans to invest in expansion over the next 6 months, 24% of them report they will fund that growth by borrowing from their main relationship bank – continuing a downward trend, and well short of the high of 38% who nominated this option to fund growth in the first round of the Index in September 2014.

21.7% of SMEs say they plan to use non-bank lenders to fund upcoming growth (with 90.8% planning to use their own funds). Non-bank lending intentions have trended upwards since the first Index, closing the gap between bank and non-bank lending intentions. Despite these intentions, more than 91% of SMEs responded in H1 2018 that in the previous 12 months they had not accessed any non-bank lending options to provide working capital for their business.

So while SMEs seem unsatisfied with traditional banks, they are not yet fully accessing opportunities available to them in the non-banking sector.

Results show that growth SMEs are five times more likely to use alternative lending options than declining growth SMEs, with debtor finance the most popular option. The growth potential for the non-bank lending sector is significant, given that 48% of SMEs who didn’t use non-bank lending in 2017 are considering it for 2018.

With SME owners revealing a solid reliance on personal credit cards to give their business the working capital required for day to day operations, those with better business solutions must find a way to reach these small business people.

Businesses implementing appropriate working capital solutions to get on top of cash flow impediments are well placed to realise their growth ambitions.

Should Central Banks Launch Digital Currencies?

The Bank for International Settlements Committee on Payments and Market Infrastructures has released a report “Central bank
digital currencies“.  It looks at both wholesale and more generally available models. The former, they say might be useful for payments but more work is needed to assess the full potential. Although a CBDC would not alter the basic mechanics of monetary policy implementation, its transmission could be affected. A general purpose CBDC could have wide-ranging implications for banks and the financial system. Customer deposits may become less stable, as deposits could more easily take flight to the central bank in times of stress.

Interest in central bank digital currencies (CBDCs) has risen in recent years. The Committee on Payments and Market Infrastructures and the Markets Committee recently completed work on CBDCs, analysing their potential implications for payment systems, monetary policy implementation and transmission as well as for the structure and stability of the financial system.

CBDC is potentially a new form of digital central bank money that can be distinguished from reserves or settlement balances held by commercial banks at central banks. There are various design choices for a CBDC, including: access (widely vs restricted); degree of anonymity (ranging from complete to none); operational availability (ranging from current opening hours to 24 hours a day and seven days a week); and interest bearing characteristics (yes or no).

Many forms of CBDC are possible, with different implications for payment systems, monetary policy transmission as well as the structure and stability of the financial system. Two main CBDC variants are analysed in this report: a wholesale and a general purpose one. The wholesale variant would limit access to a predefined group of users, while the general purpose one would be widely accessible.

Wholesale CBDCs, combined with the use of distributed ledger technology, may enhance settlement efficiency for transactions involving securities and derivatives. Currently proposed implementations for wholesale payments – designed to comply with existing central bank system requirements relating to capacity, efficiency and robustness – look broadly similar to, and not clearly superior to, existing infrastructures. While future proofs of concept may rely on different system designs, more experimentation and experience would be required before central banks can usefully and safely implement new technologies supporting a wholesale CBDC variant.

In part because cash is rapidly disappearing in their jurisdiction, some central banks are analysing a CBDC that could be made widely available to the general public and serve as an alternative safe, robust and convenient payment instrument.

In circumstances where the traditional approach to the provision of central bank money – in physical form to the general public and in digital form to banks – was altered by the disappearance of cash, the provision of CBDC could bring substantial benefits.

However, analysing whether these goals could also be achieved by other means is advisable, as CBDCs raise important questions and challenges that would need to be addressed. Most importantly, while situations differ, the benefits of a widely accessible CBDC may be limited if fast (even instant) and efficient private retail payment products are already in place or in development.

Although a general purpose CBDC might be an alternative to cash in some situations, a central bank introducing such a CBDC would have to ensure the fulfilment of anti-money laundering and counter terrorism financing (AML/CFT) requirements, as well as satisfy the public policy requirements of other supervisory and tax regimes. Furthermore, in some jurisdictions central banks may lack the legal authority to issue a CBDC, and ensuring the robust design and operation of such a system could prove to be challenging. An anonymous general purpose CBDC would raise further concerns and challenges. Although it is unlikely that such a CBDC would be considered, it would not necessarily be limited to retail payments and it could become widely used globally, including for illegal transactions. That said, compared with the current situation, a non-anonymous CBDC could allow for digital records and traces, which could improve the application of rules aimed at AML/CFT.

The introduction of a CBDC would raise fundamental issues that go far beyond payment systems and monetary policy transmission and implementation. A general purpose CBDC could give rise to higher instability of commercial bank deposit funding. Even if designed primarily with payment purposes in mind, in periods of stress a flight towards the central bank may occur on a fast and large scale, challenging commercial banks and the central bank to manage such situations. Introducing a CBDC could result in a wider presence of central banks in financial systems. This, in turn, could mean a greater role for central banks in allocating economic resources, which could entail overall economic losses should such entities be less efficient than the private sector in allocating resources. It could move central banks into uncharted territory and could also lead to greater political interference.

Voice assistants will have to build trust before we’re comfortable with them tracking us

From The Conversation.

We’re all used to targeted advertisements on the internet. But the introduction of voice assistants like Apple’s Siri and Google Assistant mean that companies are capturing all new kinds of data on us, and could build much more detailed “behaviour profiles” with which to target us.

There is already a lot of scaremongering and pushback, as there was with targeted online advertising.

But over time consumers have come to not only accept targeted advertising and personalisation, but to see it as valuable. When advertising is relevant to our interests and needs, we have the opportunity to discover new brands and products. This is a win for both consumers and brands.

A behavioural profile is a summary of a consumer’s preferences and interests based on their online behaviour. Google, Facebook and other platforms use this personalised data and activities to target advertising.

Currently these profiles are built using data on search and internet activity, what device you are using, as well as data from our photos and stated preferences on things like movies and music (among many other things).

But voice adds a whole other dimension to the kind of data that can be collected – our voice assistants could pick up conversations, know who is home, what time we cook dinner, and even our personalities through how we ask questions and what we ask about.

However, Google says its virtual assistant only listens for specific words (such as “ok Google”) and that you can delete any recordings afterwards.

Remarkably, many young consumers evidently once believed that their information wasn’t being used to target advertising at all. This 2010 study showed that even though young people knew all about tracking and social media, they were still amazed at the thought of their information being used.

The people in the study thought that if their accounts were on private then no one else had access to their information.

Targeting, good or bad?

Most of us are not fully aware of how and when our data are being collected, and we rarely bother to read privacy policies before we sign up to a new platform.

Research shows that we find the personalisation of our services and advertisements valuable, although some experts suggest that companies aren’t really using the full extent of targeting capabilities, for fear of “over personalising” the messages and customers responding negatively.

However, many of us have had the experience of having a conversation about a product or brand, only to be served up an ad for that product or brand a short time later. Some people fear that the microphones are always listening, although it is likely a coincidence.

There is even a theory in academia called Baader-Meinhof phenomenon. This is when you become aware of a brand or product and all of a sudden you start to notice that brand around you, for example in the ads. This is similar to the way that once you are in the market for a new red car, all you seem to see are shiny red cars on the road.

Baader-Meinhof theory or not, the reality is that the shift towards voice-activated search brings the potential for this information to form part of your behavioural profile. After all, if the speakers know more about you, they can cater to your needs more seamlessly than ever before.

Will we accept this use of data as readily as we accepted our online information being used to target us? Or is this new technology going to inflame our privacy concerns?

Online privacy concerns are influenced by consumers’ ability to control their information and also their perception of vulnerability. Some researchers have theorised that because speakers seem human, they need to build trust like a human would – through time and self-disclosure.

However, for many of us the benefits and rewards such as finding information in a quick and convenient manner far outweigh potential privacy concerns that result from their personal data being used.

What could be more convenient or comfortable than calling out to an all-knowing omnipresent “someone”, in the same way you might ask a quick question of your spouse or flatmate?

At the moment, these technologies are still novel enough that we notice them (for instance, when Alexa suddenly started “cackling” last week). But after some time, perhaps we will come to take this personalisation for granted, always expecting ads to be targeted to us based on what we want right now.

What it comes down to is that brands need to build trust by being transparent about how they collect data. If consumers are unsure of how that data was collected and used they are likely to reject the personalised content.

Author: Louise Kelly, Lecturer, Queensland University of Technology; Kate Letheren, Postdoctoral Research Fellow, Queensland University of Technology

Where Australia Sits On The Digital Trust Scale

Interesting article from the World Economic Forum and Harvard Business Review, comparing Digital Trust across 42 countries, based on four key dimensions. We lag on momentum it seems, well below UK, New Zealand, and USA, as well as many of the Asian and Scandinavian countries!

Australia rates quite highly on the experience and environment dimension, but low on the speed dimension.

The 2016 Sterling Flash Episode

From The Bankunderground

In the early hours of the morning of 7 October 2016, the sterling-US dollar exchange rate fell by nearly 10% within around 40 seconds. Most of this movement was reversed within the ten minutes that followed. This was one of a series of such ‘flash’ episodes in major financial markets – that is sharp and short-lived movements in price, which vastly exceed perceived changes in economic fundamentals. It certainly didn’t fail to catch the eye of policymakers, or the media.

This post summarises a recent working paper by Bank staff to understand what happened and why.

What happened?

To get at this question, we took high frequency data from the Thomson Reuters platform. Day-to-day, this platform facilitates between five and ten per cent of trading in the sterling-US dollar spot market – one of the most liquid currency pairs in the world.

Chart 1 shows data taken from this platform around the episode:

The triangles show the prices at which individual transactions took place. Those in blue (pointing down) indicate transactions initiated by a participant seeking to sell sterling. Those in green (pointing up) indicate those initiated by an order to buy.

The shaded regions show the cumulative distribution of limit orders around these prices. Limit orders are unexecuted orders to buy or sell sterling posted by prospective traders.

The relative weight of the shading on the chart shows the quantity of limit orders between a given price and the limit orders to buy/sell at the highest/lowest prices (the ‘best bid/ask prices’). As might be expected, prices further from the best bid/ask are shaded in darker colours. This indicates that there lies a larger cumulative quantity of limit orders between them and the best bid/ask.

The black line shows the midpoint – or ‘mid-price’ – between the best bid/ask prices.

From this chart we can construct a rough narrative of events:

In the minute preceding the crash, between six and seven minutes past midnight (00:06:00 and 00:07:00 British Summer Time), there was quite a large depth of orders both to buy and sell sterling (£60 million of orders in the observed ten levels of price closest to the best bid and ask prices).

But at around 00:07:00, an imbalance started to develop – with the quantity of orders to sell sterling starting to exceed those to buy.

It was at this point that a rapid succession of trades took place in sterling, at rapidly declining prices.

This imbalance became particularly severe around 00:07:17 BST. This can be seen from the large white areas in the graph, which indicate there to be (close to) no orders to buy sterling. In the half minute that followed, market functioning was severely impaired, with large ‘gaps’ in price visible between trades.

The quantity of, and balance between, limit orders to buy/sell sterling recovered after about 30 seconds but deteriorated severely again after around one minute. Shortly after 00:09 BST, there was a further sharp reduction in orders to buy sterling (corresponding to another white area on the chart).

The order book started to increase in depth around 00:09:30 BST, around 150 seconds after the initial sharp movement in price.

Thankfully, the events of the night of 7 October 2016 were without lasting consequences for financial stability, or the integrity of the functioning of the market for sterling.  Higher than usual volumes were observed during the day that followed, and measures of illiquidity (including bid-ask spreads) remained slightly elevated; but broader spill overs were generally limited.

That, however – understandably – hasn’t stopped the search for answers.

…and why?

In its report on the episode, the Bank for International Settlements (2017) found the movement in the currency pair to have resulted from a confluence of factors. These included larger-than-normal trading (predominantly selling) volumes at a typically illiquid part of the trading day. There were also sales of sterling by some market participants seeking to limit the risk associated with their positions in options markets, and to execute client orders in response to the initial fall in the exchange rate.

One important outstanding question is the degree to which the change in price witnessed during the episode was in line with the imbalance between observed orders to buy and sell.

We’d expected an imbalance in order flow during an episode like this to increase the change in price that results from a trade (or trades) of a given size. An imbalance in orders might be interpreted as indicating that some market participants were party to superior, or more up-to-date, information. This might, in turn, widen the spread at which other participants were willing to buy/sell sterling.

To assess this empirically, we develop an estimate of the change in price that is likely to result from an imbalance in orders in foreign exchange markets. This is based on previous literature. It is robust to the possibility that a large order might be split up into a series of smaller orders. This is important, because such splitting of large orders is common in foreign exchange markets, because, by transacting in smaller size, market participants can obtain a better price.

The blue bar in Chart 2 shows the range of estimates of change in price given by this model, when calibrated to past movements in the sterling dollar exchange rate.  These imply that the observed orders to sell sterling during the flash episode are consistent with a decrease in the sterling-US dollar exchange rate of between 1.03% and 2.87%, depending on the precise choice of parameters.

Chart 2: The (in)consistency between observed changes in price and those expected given observed orders to buy/sell sterling

The dots in the right-hand columns compare these estimates with the observed decline in the exchange rate. The purple triangle shows that which took place early in the episode, between 00:07:00 and 00:07:15 (roughly step (2), above). The yellow square shows the peak-to-trough fall in sterling over the entirety of the episode.

From this we can see:

  • The initial fall in sterling, during the early part of the episode, of 1.8%, is consistent with the range of estimates based on the observed imbalance of orders. This suggests that the movement in price was consistent with the arrival of a large order to sell sterling.
  • But the larger change in price that occurred between 00:07:00 to 00:07:17 cannot be explained by expected price impact of trades alone.

Was there another factor at play?

That the overall fall in sterling so vastly exceeds that predicted by our model might suggest that some some other driver might have been at play.

The report by the BIS suggests a number of other factors that might have played a role in reducing available liquidity during the episode. These include the temporary withdrawal of some market participants from their role as market makers. This dynamic may have, in part, reflected the presence of staff with lower risk limits and appetite at some institutions at that time of day. An automatic pause in trading in sterling futures contracts may also have led to a reduction of liquidity in the cash market, because some market makers are thought to rely on futures as a guide to the price at which they offer to buy/sell currency in cash markets. It is, however, difficult to know what weight to place on these different explanations.

All else equal, such a reduction in liquidity would have increased the resulting fall in price beyond that estimated to be in line with observed trading volume.

Conclusion

The events of 7 October 2016 represented one of a series of flash events occurring in electronically traded markets. No such events have, as yet, had longer lasting consequences for market functioning or stability.

Nonetheless, policymakers have recently pointed to the clear onus on central banks and the regulatory community to understand developments in these markets, and how they behave during periods of stress.

This work represents one step in that effort.

Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees