Picture caption: Joanne Atkin, Group Editor, Mortgage Finance Gazette; David Newman, Finance Director, The Mortgage Lender; David Horsley, Director of Finance, Castle Trust; Andy Hudders, Director, Whistlebrook; Gordon McMaster, Financial Services Consultant, Whistlebrook; Anthony Keeble, Chief Financial Officer, Paratus AMC Limited (Foundation Home Loans); Stephen Little, Senior Reporter, Mortgage Finance Gazette
Financial organisations will be required to follow the International Financial Reporting Standard 9 (IFRS9) which comes into force for accounting periods beginning on or after 1 January 2018. The rules are long and complicated but for the purpose of this round table we concentrated on one of the major changes, which is around the reporting expected losses.
At the moment (under International Accounting Standard 39 (IAS39)), provision for losses is on an ‘incurred loss model’ basis. The new accounting standard of IRFS9 means that in each reporting period, a lender must not only report the loans that have already gone into default; they must also assess the probability of which customers are likely to suffer arrears, what those losses will be during the lifetime of the mortgage, i.e. the expected losses, and how that impacts on the capital they must hold in provisioning for those potential losses.
Although the IFRS9 rules start in January 2018, lenders with non-calendar year ends will have up to two years after that to adhere to the new standard as David Horsley, director of finance at Castle Trust, explained: “Our reporting year end is September so IRFS9 doesn’t hit us until 2019. We are currently under IAS39 and will continue to be so for the next two financial years. September 2019 is the first financial year we will have to comply.”
It’s a similar situation for David Newman, finance director at The Mortgage Lender: “Our year end is August so we won’t have to comply until the following year either.
“Our modus operandi is that we sell our loans on origination, so we don’t hold those loans on our balance sheet. However, we do service the loans so we are going to speak very closely with the owners of those loans to discuss those provisions.
“We are completely new in the market, we only got our licence in July last year. And I am glad to say we don’t actually have any loans which are underperforming or non-performing – but I want to understand the implications of IAS39 and IFRS9.”
No historical data
For new lenders, the fact that there are no loans in arrears can actually be problematic for adhering to IFRS9.
David Horsley said: “The problem we find in firms such as ours and new start-ups is that you don’t have empirical data available in terms of incurred losses to take a view. You can’t fit it to the model so it is full of judgements, full of assumptions which the board has to sign up to. You need to find an adviser who supports you on that because they may have a different view.
“And then as you move from an incurred loss model we are going to have even more of a challenge if we don’t have the empirical data to support impairment within our book. New books of business just don’t have the loss events. And that’s the biggest challenge at the moment – referencing the data internally and externally.
“Also, in Castle Trust’s world where a significant proportion of our mortgages are second charge, you have to reference the primary mortgage for the purposes of provisioning and that is a very difficult environment.”
David Newman added: “I think the data post the new affordability stress test rules is going to be very different than for loans that were originated before that point. I think there will be a need for two models (one for older loans and one for newer loans) if you are going to look historically. There’s nothing empirical at the moment and everybody is going to have to learn what to do.”
Gordon McMaster, financial services consultant at Whistlebrook acknowledged this and said: “I think that is recognised though. You can’t be expected to have it perfect right away, it’s a learning process. Firms report the provision in their accounts, but it has to be justified so you have got to get it audited. Institutions might not have enough historical data, so they may have to make a judgement call and that is a total step change.”
What other sources of data can lenders use if they have no defaults?
David Horsley noted that credit agencies have a lot of data that can be used.
He said: “You can’t always find the data points on the whole population of your mortgage book, especially when you are a second charge lender. You can only get a certain amount of coverage and then you have to use alternative procedures to get further coverage. We could use the three credit reference agencies to get as much population coverage as we can.”
David Newman added: “If you’re a lender that credit scores and pays a rate equivalent to that credit score I think it may be easy to work out a certain level of data from the point at which the mortgage was originated. If you do that, obviously there are other credit complications, not just the score.”
How do second charge lenders get the information about the borrower’s first charge mortgage?
David Horsley explained that it all comes down to the terms and conditions your customer has with the first charge lender. When the primary charge changes it is at that point you try to gather further information about the primary mortgage.
“You need to know whether any loss events have led to payment rearrangements so you can try to extract that data from the primary lender or the credit reference agencies. The data points are there and there is a typical formula that you apply under IAS39 for impairment, which includes an emergence period. You need to identify triggers before a customer defaults; in the event of job loss, for example, that won’t necessarily trigger a mortgage default, but it is an event one can capture as part of an impairment provision.”
How do you know when a borrower loses a job?
David Horsley said: “Normally under the terms of a loan, borrowers are under obligation to tell their lender of material changes and that then becomes a public record on your credit file. You can’t force a customer to tell you if they have lost their job. But we are dependent on that data to determine an accounting act.
“There are all sorts of other triggers that you can pick up on a credit file. There may be impairment events but no losses because lender and client enter into payment arrangements and as a lender you have that flexibility.”
Anthony Keeble, chief financial officer at Paratus AMC Limited (formerly GMAC-RFC), noted that a case like that could be a big issue. “You would have to justify not putting the loan into the stage that requires the lifetime losses to be accounted for,” he said.
The IFRS9 rules state that it is assumed (referred to as a ‘rebuttable presumption’) that ‘the credit risk has increased significantly when contractual payments are more than 30 days past due’.
Anthony believes the rebuttal assumption of 30 days in arrears is ridiculous for a secure mortgage debt.
He said: “One month down may represent a significant risk on a telecoms bill or an unsecured debt that requires a change in treatment. Without all the data it is very difficult to rebut the presumption without proving it. Even though you can collect data from other sources, you are going to be challenged on matching that data to your current lending.”
Foundation Home Loans is the lending brand of Paratus AMC and was set up in February 2015 as a buy-to-let lender and last month expanded into specialist residential home loans.
Anthony said: “We have data on the GMAC back book including data on most of the loans we sold and their performance. One thing that doesn’t emerge from that is how good any of those lenders (who bought GMAC loans) were at collecting payments. All lenders’ collections processes will have been different and inherently some will have a less active approach leading to drift. It is very difficult to ascribe your capabilities around that.
“From a second charge perspective it is even more difficult. A second charge lender and a first charge lender should be incentivised to actually work together in terms of the borrower’s personal circumstances.
“Our back book has been though massive cycles and we have a huge amount of data within it so from an IFRS9 perspective we wouldn’t expect a huge amount of volatility coming through that. We didn’t originate any loans for eight years and are now an emerging lender.”
Another problem that Anthony highlighted was that if a short-term default appears in lifetime losses you have to wait effectively 13 months before it can be moved back out again, irrespective of the performance of that loan.
He said: “You have got to show a 12-month period of actual remedy before you can move it back to the earlier stage. That means you have to be absolutely sure when you put it into that stage that there is a significant event, which is why the 30-day rule is horrible. The reality is that with the forbearance protocols, lenders’ engagement with the individual borrower, etc, it is more likely to be 60-90 days anyway.”
Will organisations with simpler banking models find IFRS9 easier than the big banks?
Anthony said that organisations with one or two business lines might find it somewhat easier. He said: “The big banks will have problems where they have multiple relationships with a borrower. If a borrower impairs on their unsecured loan and they have got a mortgage, how do the banks connect all of that up and make sure of their provisioning?”
David Newman suggested creating a scale of say 1-20 based on certain events happening and it becomes more significant if a borrower moved down two or three events. He said: “There could be so many cogs within the wheel that can change an individual’s situation – it not just the arrears. People can overpay on the way back, for example, so how do you account for that?”
Anthony believes that lenders, particularly new lenders, need to get better at capturing data now. He said: “If your customer adviser on a phone call identifies the loss of job, you need to find a mechanism of recording that somewhere and then apply it to see if it has had an impact. The difficulty is making sure you get that granularity of information from a collector who is actually trying to rehabilitate a borrower, without thinking about whether a provisioning event needs to come off the back of it.”
If, or when, Bank of England base rate goes up, what impact would that have on reporting expected losses?
Andy Hudders, director at Whistlebrook, recalled a conversation with a building society chief executive who said that whilst Brexit negotiations were going on and the uncertainties surrounding Brexit, the Bank of England would not allow base rate to rise and was of the opinion that it would stay put for at least two years.
David Newman thinks that Brexit could potentially be a trigger event, as could another independence vote in Scotland. He pointed to Aberdeen (where the housing market had been impacted by the drop in oil prices and subsequent job losses within the oil industry) and asked: “Do you then factor that into your calculations for every property in that area?”
If a whole work force lose their jobs, for example, the demise of BHS, you could technically go down to that level to assess potential losses.
Is IFRS9 labour intensive for organisations?
Anthony commented: “From an analytical perspective it has become much more labour intensive. Some of the bigger banks have got teams of hundreds of people working on this. Part of this is the breadth of lending and the intertwined nature of lending to individual borrowers. They are under pressure from a regulatory perspective as well as a markets perspective and accounting standards perspective to make sure they get this right. So they are under scrutiny from pretty much everyone.”
David Horsley also noted that there are other elements such as the senior manager’s regime coming in: “The expectations on managers and directors within an organisation when it comes to things like accounting are going to be even more relevant. I think the population of directors that are capable of understanding this is getting narrower.”
Do lenders have to run with the IAS39 and IFRS9 in parallel?
Anthony said that lenders don’t have to run the two standards in parallel but the auditors will recommend you do so for a period of time.
He commented: “That way you can see what the new standard does to your numbers and how close they are to your historic numbers; and actually are you seeing a huge amount of volatility as a result of the new standard? But I would guarantee all the big banks have been at least running some version of IFRS9 for many months and refining it.”