Bank launches AI ‘home loan assistant’

From The Adviser.

A new chatbot has been launched by UBank to help answer customers’ home loan questions, in what is being billed as Australia’s “first virtual assistant for home loan applications”.

RoboChat has been launched by the NAB-owned online bank as a “virtual assistant to help potential homebuyers and refinancers complete their online home loan applications” and “simplify the home loan application process”.

The new chatbot has been built by IBM Watson to act as a virtual assistant to aid customers through the home loan application form.

According to UBank, the bot — which is still “in training” — aims to “help reduce the time needed for customers to complete the form by offering on-the-spot help”.

It has been trained on data collected from customer questions submitted via UBank’s LiveChat experience, and has been tested by “dozens of users and iteratively trained”.

UBank has said that the bot’s artificial intelligence will “continue to learn as more customers engage with it, becoming smarter and more user-friendly over time”.

However, the bank has been emphatic on the fact that the bot “won’t affect the size of the local UBank customer service team or the great work they do”.

Speaking of the product, the CEO of UBank, Lee Hatton told The Adviser that RoboChat was created to help customers overcome the “fear of how long it will take to fill out the applications”.

When asked if the platform was meant to emulate what a broker does, Ms Hatton said: “RoboChat was designed specifically to guide customers through UBank’s online home loan application form.

“UBank doesn’t use mortgage brokers, so RoboChat, along with our experienced advisers, can help customers as they decide which home loan is right for them.”

Ms Hatton added that the chatbot is available 24 hours a day, 7 days a week so can provide real-time answers, but customers can still use a bank adviser to assist with “more specific support or [questions] fitting to [an] individual situation”.

Putting RoboChat to the test

Once a user selects that it wants to start a home loan application, the RoboChat option opens up alongside the online application process. It is here that users can ask the bot questions. The website notes that once an application is completed (with questions answered), a home loan specialist will be in touch to move the process forward.

When first opened, the chatbot introduces itself as Ubank’s “beta virtual assistant”, adding that it is “still in training” and will “do [its] best to answer your home loan questions”. It says you can speak to a real person at any time.

The Adviser tested out the platform, and it successfully answered simple questions such as “What is your variable rate?’ (with that answer being 3.74 per cent) and “Do you have an offset?” (which it doesn’t), as well as “What are your charges?” (it doesn’t have any unless the loan is fixed rate) but struggled when the next question was a new subject (“Can I have an interest-only loan?”), responding that it was “wasn’t expecting that kind of response”.

However, once establishing that the user was still interested in interest rates, and ascertaining that the user was an owner-occupier, the bot asked if it was for principal and interest, or interest-only. Once interest-only was chosen, the bot produced the interest rates with comparison rates (4.13 per cent with the Real Reward Offer or 4.46 (4.39 per cent comparison) without).

When asked “Should I use a broker?” the RoboChat asked for the question to be rephrased. When asked “Do you have a broker?”, it said: “We’re an online bank so we don’t have branches, mobile lenders, or use brokers. The upside is we have fewer costs than traditional banks and we pass the savings onto you. Do you want to know more?”

It also understood “Can you tell me your LVR?”, but struggled with “What is your LVR?”, and didn’t understand “Can I refinance?” but knew the answer to “How do I refinance?”.

When asked “How much can I borrow?” RoboChat said: “I can’t perform calculations just yet” and offered links to tools and calculators.

The platform handled other questions, such as “I’m self-employed. Do you have a loan for me?”, responding “To ensure we’re a good match, your primary income needs to be from an Australian PAYG source of a company not owned by you. And, all applicants must be either an Australian citizen OR Australian Permanent Resident OR a NZ citizen AND reside in Australia. As a bank, we’ll always need to perform a credit assessment based on the details of each applicant.”

While the bank has said that it has in-built humour and offers “tongue-in-cheek responses”, the bot was not able to respond to the joke as outlined in the press release (“How much does a hipster weigh?” Answer: “An Instagram”), instead saying “Hmmmm, I’m a bit confused. Can you rephrase your question?” When asked for the meaning of life, it did say, however, “Simpler, better, smarter is my life motto!”

Ms Hatten commented: “Our goal is to deliver simpler, better, smarter banking to our customers and RoboChat will help deliver on this by streamlining the application form.

“RoboChat will be a very welcome addition to our team of customer service experts” continued Hatton.

“UBank will still have experienced staff on hand to chat on the phone, via email and our live online chat offering, RoboChat will provide an added option for those needing quick online responses or those that are close to finalising the form.”

Brock Douglas, vice president of Watson for IBM Asia Pacific added: “From deepening the customer experience, to increased productivity for employees, virtual assistants are being adopted across industries and becoming more advanced in natural conversation and emotional intelligence, with the help of cognitive technology.

“UBank’s work with IBM Watson is a powerful example of how organisations are leveraging cognitive virtual assistants that have the ability to engage in a conversation, ask questions, learn and respond in context – as opposed to providing stock responses.”

Bank Stress Tests Are Not Up To The Job

An important IMF working paper, released today suggests that the standard stress test models used to assess risks in the banking system are likely to underestimate the impact of stress on bank solvency and financial stability because they do not consider the dynamics between solvency and funding costs.

The global financial crisis appears to have been a liquidity crisis, not just a solvency crisis. Yet the failure to adequately model interlinkages and the nexus between solvency risk and liquidity risk led to a dramatic underestimation of risks. Liquidity risk manifests primarily through a liquidity crunch as firms’ access to funding markets is impaired, or a pricing crunch, as lenders are unwilling to lend unless they receive much higher spreads.

A sudden increase in bank funding costs can have an adverse impact on financial stability through the depletion of banks’ capital buffers. To preserve financial stability, it is important to assess banks’ vulnerability to changes in funding costs. The reason is twofold. First, to the extent funding costs reflect counterparty credit risk, it is of particular interest for supervisors to determine the level of capital buffers that should be held to keep funding costs at bay if and when market conditions deteriorate. Second, funding costs are linked not only to banks’ initial capital position but also they determine their capital position going forward, paving the way for adverse dynamics. The magnitude of this effect is likely to depend on the bank’s behavioral reaction to rising funding costs. On the one hand, it may react by setting higher lending rates to its borrowers. Yet this action reduces the bank’s market share and its franchise value. On the other hand, the bank might not be able to passthrough additional funding costs to new lending so its internal capital generation capacity is reduced. Even if some pass-through is possible, the erosion of profits is likely to be
substantial given the shorter time to repricing of liabilities relative to assets with the margin impact on the carrying values of assets outweighing that of new asset generation.

“Bank Solvency and Funding Cost: New Data and New Results”  presents new evidence on the empirical relationship between bank solvency and funding costs. Building on a newly constructed dataset drawing on supervisory data for 54 large banks from six advanced countries over 2004–2013, we use a simultaneous equation approach to estimate the contemporaneous interaction between solvency and liquidity. Our results show that liquidity and solvency interactions can be more material than suggested by the existing empirical literature. A 100 bps increase in regulatory capital ratios is associated with a decrease of bank funding costs of about 105 bps. A 100 bps increase in funding costs reduces regulatory capital buffers by 32 bps. We also find evidence of non-linear effects between solvency and funding costs. Understanding the impact of solvency on funding costs is particularly relevant for stress testing. Our analysis suggests that neglecting the dynamic features of the solvency-liquidity nexus in the 2014 EU-wide stress test could have led to a significant underestimation of the impact of stress on bank capital ratios.

The results are also highly relevant for cost impact assessments of capital regulation, as the costs of higher capital requirements are partly offset by lower debt servicing costs.

Note: IMF Working Papers describe research in progress by the author(s) and are published to elicit comments and to encourage debate. The views expressed in IMF Working Papers are those of the author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.

US Jobs Up

The US Bureau of Labor Statistics says the March 2017 job openings rate for total nonfarm was 3.8 percent, and the hires rate was 3.6 percent. Job openings rates in finance and insurance; professional and business services; health care and social assistance; and accommodation and food services were higher than the overall rate.  Rates were lower in mining and logging; construction; durable and nondurable goods manufacturing; wholesale and retail trade; transportation, warehousing, and utilities; information; real estate and rental leasing; educational services; arts, entertainment, and recreation; and federal, and state and local government.

Hires rates were higher than the total nonfarm rate in mining and logging; construction; retail trade; professional and business services; arts, entertainment, and recreation; and accommodation and food services. Hires rates were lower in durable and nondurable goods manufacturing; wholesale and retail trade; transportation, warehousing, and utilities; information; finance and insurance; real estate and rental leasing; educational services; other services; and federal and state and local government.

Industries with high job openings rates and high hires rates need more workers, and hiring is strong. In March 2017, industries in this category included professional and business services and accommodation and food services.

Industries with low job openings rates and low hires rates have few job openings and are hiring few workers. In March 2017, industries in this group included federal and state and local government, educational services, and information.

The data is preliminary, from the Job Openings and Labor Turnover Survey and are seasonally adjusted.

Even God now accepts digital payments

From Bloomberg.

In the most cashless society on the planet, even God now accepts digital payments.

A growing number of Swedish parishes have started taking donations via mobile apps. Uppsala’s 13th-century cathedral also accepts credit cards.

The churches’ drive to keep up with the times is the latest sign of Sweden’s rapid shift to a world without notes and coins. Most of the country’s bank branches have stopped handling cash; some shops and museums now only accept plastic; and even Stockholm’s homeless have started accepting cards as payment for their magazine. Go to a flea market, and the seller is more likely to ask to be paid via Sweden’s popular Swish app than with cash.

“Fifteen years ago I would withdraw my entire salary and put it in my wallet, so I knew how much I had left, but these days I never really carry cash,” said Lasse Svard, the acting vicar at the parish of Jarna-Vardinge, about 50 kilometers (31 miles) south of Stockholm.

Vanishing Act

Swedes’ aversion to cash is increasingly showing up in money supply data. According to Statistics Sweden, notes and coins in public circulation dropped to an average of 56.8 billion kronor ($6.4 billion) in the first quarter of this year. That was the lowest level since 1990 and more than 40 percent below its 2007 peak with the pace of the decline accelerating to its fastest ever in 2016.

According to the central bank, which is also studying whether to launch its own digital currency, the main reason for the disappearing act is technical innovation.

Riksbank Deputy Governor Cecilia Skingsley notes that Swedes were early adopters of both personal computers and mobile phones (remember those Ericsson phones?), and that the country’s banks were quick to create sector-wide structures such as debit cards, credit cards and Swish, which has 5.5 million users and is owned by the country’s largest banks. Swedes also seem to trust those systems, she said in a recent interview in Stockholm.

Innovation Drive

“A drive for innovation has been created in Sweden to come up with cost-effective and user-friendly alternatives to cash,” Skingsley said. Cash is likely to “more or less disappear” as a means of payment in the private sector, she said.

 

But a cashless society is not without its challenges or critics.

Many pensioners are struggling to make payments in an online world, while privacy campaigners lament the fact that the state is acquiring greater control over what its citizens do. There are also concerns about the vulnerability of a cashless society in the event of an attack or major blackouts.

It seems that for now, the benefits — including lower business costs, more control over tax revenue and greater safety from criminals — outweigh the drawbacks.

Policy Impact

In fact, Swedes have become so averse to cash that they’ve even been shunning it during the current regime of negative interest rates.

“The fact that people chose to hold so little cash in their wallets despite getting zero interest on their bank accounts emphasizes the strength of this trend even more,” Robert Bergqvist, chief economist at SEB AB, said by phone.

So far, the development has had little impact on monetary policy. But if cash were to disappear entirely, it would give extra powers to the central bank.

“Negative interest rates would be more powerful as there wouldn’t be any way to escape their impact,” Bergqvist said. “In a sense it would give the Riksbank more room for maneuver.”

 

Banks seek to delay levy implementation

From InvestorDaily.

The Australian Bankers’ Association (ABA) has called for the “usual consultation and policy processes” prior to the implementation of the bank levy contained in last Tuesday’s federal budget.

In a submission to Treasury on the major bank levy, the ABA said the budgetary measure is “rushed” and “not in keeping with the government’s own best practice guidelines”.

The levy would impose a tax of 6 basis points on the assessed liabilities of the ‘big four’ and Macquarie.

In its submission, the ABA called on Treasury to conduct a detailed regulatory impact status before the bank levy bill is introduced to Parliament.

In addition, the ABA repeated its calls for increased modelling on the economic and taxation impacts of the proposed levy.

The bank lobby group also pointed to the short consultation period on the draft legislation, which is set to be finalised before the bill is introduced to Parliament on 31 May 2017.

“The ABA are alarmed with the truncated time for consultation, as well as the fact there was no prior consultation, nor will exposure draft legislation be released for public comment,” said the submission.

“The ABA believes there is further opportunity for consultation as the tax will only be levied for the first time on 30 September 2017.”

The ABA also argued for a more “co-ordinated consultation” with all the affected regulators, including APRA, ASIC, the AOFM and the RBA.

“For example, the ABA believes it is crucial that Treasury and APRA be given adequate time to assess if the bank levy is consistent with developments in prudential regulation such as the unquestionably strong requirements and Total Loss Absorbing Capital (TLAC),” said the submission.

Why rent bidding apps will make the rental market even more unaffordable

From The Conversation.

Renters are already the weak party at the negotiating table because they cannot ignore their need for a place to live. But a new series of apps that pit renter against renter will only further tip the balance of power in favour of landlords, making it even harder to get a house.

The fierce competition in the rental market often results in renters paying more than is necessary. Over 150,000 tenants pay more than 50% of their incomes for housing.

These apps may seem like they give renters more power – they are marketed using words like “fairness” and “transparency”, but they also note that landlords are missing out on “millions and millions”.

Renting is a zero-sum game. Every dollar that a landlord gains is a dollar out of the tenant’s pocket. And in a market already tilted in favour of landlords, these apps could further push up rents.

To address the problem of renting affordability we need technologies that promote more cooperation between renters, rather than competition.

The apps

There are several different rent bidding apps, and they all work in different ways.

Live Offer asks prospective tenants to fill out forms and these are then ranked for the landlord to choose. The prospective tenants can see where they are in the rankings, in real time.

Rentberry is more of a real-time auction site. Prospective tenants submit bids, can see what the current highest bid is and how many bids there have been.

With Rentwolf, prospective tenants set up extensive profiles, as they would with AirBnB, and then apply for properties through the marketplace.

What is the right price?

Real estate is worth what people are prepared to pay for it. But tenants will always be the weak party at the negotiating table.

At least as far back as David Ricardo in 1817, economists have theorised that landlords take what is left over once other costs are deducted from the tenant’s income. In other words, rent is as much as tenants are able to pay.

Even before that, Adam Smith recognised that all real estate is a monopoly for landowners. This is because the supply of land is strictly limited, giving excessive negotiating power to whoever owns it.

But these apps overwhelmingly rely on auctions, which can itself be a problem due to what is called the “winner’s curse”.

Studies have shown that so long as there are at least two motivated bidders, the winning bid tends to either equal the value, or be one bid above it. In other words, the winning bidder in an auction will often overpay and will “suffer” for having won.

In the case of renters, this means paying excessive rents throughout the lease. Even worse, as more people bid similarly, the ruinous rents become accepted as normal. They become the market rent.

This same phenomenon has been shown in everything from jars of coins to oil-drilling rights and, yes, real estate.

Some research has further shown that in an auction the second-highest bid could be the “rational price”.

My own controlled experiments have shown that people are easily encouraged by necessity to bid excessively in auctions. This is despite full knowledge of the ruinous consequences.

This is because people have only a limited capacity to vary their needs. We all need to live somewhere, and our culture limits the options. If the option is sharing with relatives or accepting a lower standard of living due to high rents, then our sense of independence will often prompt our willingness to tighten our belts.

These rental apps will play into these stresses and uncertainties, making it more likely people will overbid in the auctions. This is why these rental apps are likely to result in higher rents.

How technology can reduce rents

Some time ago researchers suggested that renters might be able to reduce rents by banding together.

Working together, renters would be able to create the equivalent of there being only one buyer in the market – a monopsony. A monopsony works like a monopoly, but for buyers rather than sellers. For real estate rents, if land owners enjoy a natural monopoly-like advantage, then tenants have to behave like a monopsony to negate the power imbalance.

The current crop of rental apps do not address this imbalance. They only inform bidders as to what other, similarly stressed people, are bidding. It stresses them to bid higher.

Already, we are seeing the power of Facebook to build communities around renting. There are groups that help people find flatmates, for example. If the internet is to tackle the problem of renting affordability, then we need to extend this community, for renters to act in unison on rent.

Without a service that seeks to unite renters, rather than have them compete, housing affordability will only get worse. Rent bidding apps will increase landlord revenues and do so at the expense of tenants.

Fewer Buyers Results in Greater Months of Supply

From CoreLogic.

Based on the relationship between demonstrated housing demand and advertised stock levels we are seeing relatively more stock available for sale compared to demand for that stock across the capital cities at the moment.

The months of supply figure compares the number of unique properties advertised for sale to the number of transactions in the market.  The analysis provides unique insight into how long it should take to clear the volume of stock currently available for sale.  It is important to note that off-the-plan housing stock is typically not advertised for sale as individual properties and as a result is not included within this analysis.

Across the combined capital cities, there is currently 4.4 months’ worth of residential properties being advertised for sale.  As the first chart shows, the months of supply is currently higher at this point of the year than it has been each year since 2012.  The increasing months of supply is largely due to the slowing rate of transaction activity rather than a spike in properties available for sale, indicating demand has diminished relative to advertised supply levels which is pushing the figure higher.

Months of supply, combined capital cities

2017-05-15--image1

Sydney

There is currently 3.0 months of established housing supply available for sale across the city.  The city has seen consistently low months of supply over recent years however, the figure is currently at its highest level for this time of year since 2012.

2017-05-15--sydneyimg

Melbourne

Established housing stock currently sits at 4.2 months of supply for the city.  Supply levels remain quite low but are at their highest levels for this time of year in five years.

2017-05-15--melb

Treasury Receives Big Bank Feedback

NAB and ANZ have released their formal responses to Treasury relating to the liability tax.

NAB believes the levy is poor policy and, accordingly, does not support it. It says the levy is not just on banks, it is a tax on every Australian who benefits from, and is part of, the banking industry.  They then try to define the bounds and sensitivity of the calculation, with a view to reducing its impact.

NAB requests the production of a Regulatory Impact Statement (RIS) and a period of public consultation on the draft legislation. NAB recommends the levy be applied to the netted derivative balance sheet and collateral position. NAB recommends that the basis for the levy be adjusted for the impacts of the accounting gross ups which occur as a function of inter-company transactions. NAB recommends that the funding of high quality liquid assets be excluded from the levy calculation. NAB recommends that repurchase agreements be excluded from the calculation of the levy. NAB recommends the exclusion of non-funding liabilities, in particular, liabilities and provision for taxes and the levy. NAB recommends that, if included, only targeted anti-avoidance measures are contained in the
legislation. NAB also recommends that discretion be applied on any penalties for under payment.

It needs to be read in conjunction with their CEO’s earlier comments.

ANZ has more broadly tried to explain the potential impact of the tax on customers and shareholders. They also suggest a delay till September 2017 to allow sufficient time for design of the legislation and recommends the tax should be applied to the domestic liabilities of all banks operating in Australia with global liabilities above $100 billion. Finally, they argue the levy means it would be appropriate to re-think the need for any bank loss-absorption framework in Australia.

Westpac said last week:

“Westpac Group CEO, Brian Hartzer, said the new bank tax is a hit on the retirement savings of millions of Australians as well as all bank customers.

This levy is a stealth tax on their life savings, the shares in their superannuation accounts, and it will make Australia’s banks less competitive.

“Yesterday, $14 billion of value was wiped off Australian bank shares because of speculation around this new tax.

“There is no ‘magic pudding’. The cost of any new tax is ultimately borne by shareholders, borrowers, depositors, and employees.

“The Australian banks are already the largest taxpayers, with Westpac the country’s second largest taxpayer. Westpac already pays over 30% of its profits in tax and this will now increase even further,” Mr Hartzer said.

“While similar taxes operate in other international jurisdictions, they were introduced to recover the cost of Governments having to take over their banks. No taxpayer funds have been used to prop-up the Australian banks. In addition, international jurisdictions that apply measures such as this already have much lower corporate tax rates than Australia – for example, in the UK the corporate tax rate is 20%.

“It is disappointing that the Australian Government has implicitly favoured large foreign banks over Australian banks operating in their home market.

“In addition these reforms are directly counter to APRA’s objective of making the banks unquestionably strong, as higher taxes reduce the banks’ ability to generate capital that supports lending and stability in times of stress.”

Housing Finance On The Slide

Latest data from the ABS for March 2017 shows that the trend estimate for the total value of dwelling finance commitments excluding alterations and additions was $33.4 billion, down 0.1%. Owner occupied housing commitments was $20.1 billion up 0.1% while investment housing commitments was $13.2 billion down 0.3%.

Within the mix, owner occupied refinance flows fell 1.1% and investment finance by individuals fell 0.5%.  In trend terms, the number of commitments for the construction of dwellings rose 0.4% and the number of commitments for the purchase of new dwellings rose 0.2%. The number of commitments for the purchase of established dwellings fell 0.1%.

Investment housing lending comprised 40% of monthly flows, refinance 18.4% (and still trending down) whilst funding for new construction continued at about 8.7%.

Total ADI loan stock grew by 0.4%, with owner occupied loans growing just a little faster than investment loans.

Investment loans comprise 35% of the total book.

The number of owner occupied first time buyers rose in March by 20.5% to 7946 in original terms, a rise of 1,350.  In original terms, the number of first home buyer commitments as a percentage of total owner occupied housing finance commitments rose to 13.6% in March 2017 from 13.3% in February 2017.

The DFA surveys saw a small rise in first time buyers going to the investment sector as their first property purchase. Total first time buyers were up 12.3% to 12,756, still well below their peak from 2011 when they comprised more than 30% of transactions.

Note the ABS has revised the data series:

In this issue, revisions have been made to the original series as a result of improved reporting of survey and administrative data. These revisions have affected the following series:

  • Owner-occupied finance from May 2012 to February 2017.
  • Investment housing finance from July 2013 to February 2017.
  • Housing loan outstandings to households for owner occupation series for the periods May 2016 to January 2017.

The number and value of commitments for the purchase of newly erected dwellings have been revised for the period May 2012 to June 2013, and consequentially the number and value of commitments for the purchase of established dwellings have been revised for the same period.

 

Auction Volumes Rebound

From CoreLogic.

Auction markets have remained resilient, with both volumes and the preliminary clearance rate rising week-on-week. The strong auction results add some complexity to speculation that the housing market is moving through its peak rate of growth.  At face value, auction markets are continuing to indicate continued strength in selling conditions across Sydney and Melbourne, however it’s harder to know whether vendors are adjusting their reserve pricing in order to clear their property. There were 2,376 auctions held across the cities this week, with a preliminary auction clearance rate of 76.2 per cent. Last week, a final clearance rate of 73.0 per cent was recorded across 1,689 auctions. Over the corresponding week last year, auction volumes were lower than this week, with 1,876 properties taken to auction and a clearance rate of 69.5 per cent. Melbourne had the highest number of auctions this week, with 1,092 properties going to market, with a lower preliminary rate of clearance week-on-week (76.8 per cent), however Sydney saw the largest increase in volumes over the week, with 938 auctions held across the city, increasing from last week’s 592.  Final auction clearance rates are published on Thursday and it will be important to monitor whether preliminary clearance rates undergo some revision as more data flows through.