There is a heck of a lot in this, so I have split this into two posts – one on the exposure draft (ED) on amortised cost (AC) and impairment and one on the (now released) “final” standard on classification and measurement – which is also long and will take a day or so. If you want to know why that is in scare quotes, go there when it appears.
As you can tell from the title, this is the one on amortised cost and impairment.
To read what the IASB has to say on this, go here for the press release and here for the content. As previously advised, I would suggest sticking to the basis for conclusions (BCs) as they are more readable, but if you want to submit a comment, you will need the exposure draft.
Having now read this a couple of times and listened to the webcast, I am still uncertain in a number of areas. I think this is going to take some work over the next couple of weeks to get sorted out and then the (now) three years’ implementation period to get roughly right.
That’s right – three years. As part of the release of this section the Board said that the earliest date for mandatory adoption will be 1 January 2013 – with early adoption permitted. This is apparently as they realise the complexity of this (I feel like saying no and a certain four letter word at this point) and to align with the prospective new insurance contracts standard. It may even be delayed past 2013 if that standard is delayed. In the webcast to night they were at pains to say they did not want to have insurance companies unduly inconvenienced by having a double change of standards.
I will work through the ED in the order in the BCs as I have advised reading those. Where there is something in the ED that I find interesting I will cover it even if it is not mentioned in the BCs.
Scope
This ED does not try to revisit the scope of IAS 39. They recognise that there may be some problems, but, in essence, they have put that in he “too hard” basket at this stage. They may have a go at it in the final standard, but not here.
Personally, I doubt they need to. The current scope, although not without some issues, is clear enough. The only issue I have ever come across with the scope statement is whether gold is counted as a financial instrument or a commodity. While others here may argue that it should be money (you know who you are), I would say this is irrelevant as it is clearly a commodity in the way it is used at the moment.
Impairment
This is the biggie here. The current IAS 39 incurred loss model is consistent with the one in the actual impairment standard, so a change here (as is mooted) would mean that financial instruments are treated differently to everything else. After some too-ing and fro-ing, they seem to have decided this is warranted and suggested plumping for an expected loss model – but using a point in time (Basel II style) estimate. This would then mean that the old “Allowance for Doubtful Debts” approach would be re-born. Talk about back to the future.
They then get into some really tortured language on the issues with it and why they did this. It actually goes on for several pages noting some of the difficulties.
I recognise these difficulties and, having worked in both Basel II Advanced institutions and in several small non-financial institutions, I can only see this approach and all the verbage around it as likely to cause confusion. As with several other areas in this ED, it looks like this has been written as if it will only apply to banks and other financial institutions – and big ones at that. This proposed approach looks like a nightmare for any one else, and difficult as well for banks, using as it does the PIT rather than the more common TTC method.
Even with three years to implement I can only see this as being well-nigh impossible for a normal company to implement. Even big non-banks will have serious problems. I suspect this is going to be a fruitful source of work for me over the next few years.
The use of an expected loss model also opens up the financial
statements to game playing by management, as they can directly
manipulate the financial statements by changing their own ideas about
the future to suit the outcome they want. Provided these expectations
are not completely outlandish they are not really susceptible to audit. I
see this as a possibly retrograde step. While I recognise the problems
with an incurred loss model I am not sure that abandoning it is
appropriate, as I said in my previous piece on this. Let’s say I am unconvinced at the moment.
Objective Statement
Motherhood statements in an accounting standard.
Measurement Principles
Ditto.
Presentation
In this section and the next we get into the likely consequential changes to IFRS 7 arising from IFRS 9.
Having just re-iterated the measurement principles, the IASB is now saying that these only apply to the finalised numbers – you also need to present, on the face of the financial statements, the interest and expected losses prior to any discounting. The effects of changes in estimates will also be there.
I can understand these being in there – they are useful numbers for analysts – but I am not sure we need to clog up the statements in this way. That said, given the next section, putting them here may mean they are not lost.
Disclosure
As could probably have been expected, there is an awful lot more in the disclosure section. The new impairment model, if adopted, would need a raft of new disclosures. The new (old) allowance account needs comprehensive disclosures including a reconciliation of changes in this account. There is also a raft of new disclosures around changes in estimates – including “reasonably possible alternative assumptions” and estimation techniques.
The devil here will be in the detail. I will need to go through these properly and see what is there. Some of these look like they could form books in themselves if they go forward as they are. Again – these look calibrated to financial institutions, rather than any other form of company.
The disclosures on any stress testing are interesting – if you are doing it you would need to disclose the fact and the results.
This seems, at best, questionable. Financial institutions are already doing this and having to disclose this through their Basel II Pillar 3 reporting. For any other company the fact they would have to make these sorts of disclosures if they conduct stress testing may just put them off doing it – a very perverse outcome as stress testing is a very useful risk management tool.
Credit Quality
The good thing here is that it looks like the IASB has picked up on the move to 90 days past due or non-collectable as the key non-performing indicator.
This looks sensible. BTW – can someone please tell APRA that his is not 4 missed payments, but three?
Vintage Information
No – not winery data, but a new grid (see B29, p33 of the ED) showing the origination and maturity date for any AC products held – and at nominal amounts.
I hope they will not expect a reconciliation here. For a bank with some products going out, say, 30 years this on will be a nightmare, as well as difficult to fit on a page. I hope the final version will allow maturity buckets. For non-banks this will merely be difficult, but not impossible to fit on a page for most. I wonder if we will get fold outs, a-la Playboy centrefolds, for this one from a bank with maturities out 30, or even 50, years.
Transition
I have read this a couple of times. As noted above, they realise that not many are likely to early adopt. they also note that the early adoption of a “superior” impairment model is likely to produce benefits. As for transition itself, apart from the bits where they say they rejected both full prospective and full retrospective implementation I still am not entirely sure what they mean. In BC 75 they outline a possible change process, that of adjusting the EIR to the new method, but I am going to have to re-read this when I have had a good think about the likely consequences.
Alternative View
The last two pages of the BC are worth a read – and they are pages I have much personal sympathy with. Two members of the IASB do not agree with the proposed expected loss method and want to stick with an incurred loss model.
As above, I have some sympathy to this view.
Summary
An enormous amount of work coming up – and over three years. The big four, who are the only ones likely to come close to getting this right as going to be looking at their fee base and realising that the lack of hiring over the last couple of years was a bit of a mistake.
Many of the disclosures also look like the IASB has gone for a “me too” approach to Basel II Pillar 3 – mandating the sorts of disclosures for all that are currently only made by banks. This I find odd, as the banks are already making them and I am not sure what use they are for readers of the reports of non-banks.
This was also only a quick look and I have focussed on the areas I believe to be of most concern. Feel free to point out others in comments.
5 comments
16 November, 2009 at 22:49
bruce
Thanks for the post Andrew. Helpful.
You obviously read the text more thoroughly than me!
I do have a slight issue with the existing incurred loss model as some stretching is required with the recognition of loss evidence. Ie a proportion of exposures will become impaired in the near future although no current evidence exists. The IAS39 rules allow for this to some extent but it is contrary to the key aims of the incurred loss approach.
However, with the new rules, it seems the IAS is trying to duplicate some of the principals of capital (although not going is far as the Spanish system of counter-cyclical provisions).
The mention of stress testing is interesting. For me, the key goal for Basel Pillar 2 anaylsis is stress testing the Pillar 1 capital requirements as defined by the IRB models and also the likely errosion of capital during the stress scenario due to impairment charges and other P&L costs.
However, if there was a requirement to publish this sort of analysis in detail, there could be a process of “window dressing” involved rather than realistic shock testing.
Will need to read it again. a few times…..
17 November, 2009 at 09:23
Andrew
bruce,
I would agree that the incurred loss model has its problems, one of which relates to the IBNR-like losses, but these are (or at least should be) a small part of the portfolio. Once you throw in expected losses, however, they would be drowned out in the tidal wave of losses that could be added in, essentially on management whim – with words like “prudence” and “conservatism” used as cover.
I found the incurred loss model difficult to implement when it first came in – and I had to be one of the first in the country to implement it – but I believe it is likely to be more correct than the previous EL model, which had always been wide open to management manipulation. The previous EL model was always a nonsense – using arbitrary numbers to push profits up or down largely at whim.
Perhaps the IASB can nail an EL model down – but I cannot see how a nailed down one is then going to be simple to implement for anyone other than large banks that are already modelling EL for Basel II purposes.
On stress testing I would agree. The inclusion here is dangerous and likely to result, as you say, in window dressing rather than real useful information.
The more I think about it the more I believe that we need to have an incurred loss model for financial reporting and leave the EL models to the prudential reporting.
17 November, 2009 at 11:59
nzriskmanager.com » Blog Archive » Loss provisions
[…] will be keeping a close eye on ozrisks blog who have been analysing the revision of IAS39/IFRS 9 Related postsNo Related Post (No […]
20 November, 2009 at 04:10
bruce
One aspect that interests me is the idea of clearly disclosing the priced-in interest margin to cover losses. This appears to be part of the proposal but not using a simple, clear method.
I do think this could be a useful disclosure for analysts to judge how sustainable a lenders business model is. These would cover the baseline costs due to losses for a reporting period and could be compared/ balanced against the known (incured) losses and actual writeoffs for each reporting period. Any material changes in loss experience (outwith cyclical behaviour) would be reported as an adjustment to the priced-in margin and would result in a cost/charge.
20 November, 2009 at 19:07
Andrew
It would be an interesting disclosure for financial institutions. The problem is that this standard applies to all companies (at least in Australia) and financial institutions are already making similar (although not this particular one) disclosures under Basel II.
Personally I think that this one will attract a lot of bad press (from, at least, the financial institutions) and be dropped from the final standard. It is also dependent on the re-introduction of the expected loss model – one of the hesitations I have with this ED.
As I get more familiar with these disclosures I just wonder about how much is enough. I would expect a number of the submissions to “suggest” that the quantity be cut down – or at least some disclosures restricted to particular types of companies.