Friday, November 1, 2013

DoF Mortgage Arrears Release

Yesterday, the Department of Finance published the first in a new set of monthly mortgage arrears and restructures figures.  Hopefully, the series will be expanded because the first issue contains almost nothing that is new and also has some errors.

The errors don’t relate to the arrears figure but to the comparison between the number of houses in the country and the number of mortgage accounts.  This is shown in the ‘key highlights’ (click to enlarge):

Mortgage Restructures

The left panel of the middle section indicates 700,000 of the 1.9 million houses in the Ireland have “mortgages covered under MART”.   This is not correct.  Yes, there are nearly 2 million housing units in Ireland, as was measured with Census 2011:

Status of Occupancy

However, the number of units with a mortgage is not the same as the number of mortgage accounts.  A separate measure in the Census showed that of the 1.65 million units occupied by the usual residents that  just over 580,000 were owner-occupiers with a mortgage.

Nature of Occupancy

The 699,674 used in the DoF release would only be appropriate if the relation between mortgage accounts and houses was 1:1. It is not.  There are many household who have more than one mortgage account on the principal dwelling house (PHD).  This can be because of top-ups, remortgages or splitting a mortgage between different interest rate types. 

Previous work by the Central Bank has shown that the ratio of mortgage accounts to PDHs is around 1.27:1.  Therefore the 699,674 mortgage accounts in the MART corresponds to roughly 550,000 houses/households.  This means 27.6% of houses in Ireland have “mortgages covered under MART” not 35%.

The figures on mortgage restructures add little to what is already known from the quarterly figures published by the Financial Regulators office.  There is a breakdown of “permanent” versus “temporary” but this is broadly known from the type of restructure used.  It would be useful if this breakdown was further distilled into those accounts which are in arrears and those accounts which are not.

It would be even more useful if the success rates of the restructures were published, i.e. whether the borrowers are meeting the terms of the restructures.  The figures from the FR indicate that 76% of the restructures are being adhered to but we do not know which restructures this applies to.

The fact the 75% of mortgages in 90 day arrears or more have not been restructured is not news but it did make the front pages of today’s Irish Examiner (image) and The Irish Times (image).

Most mortgages 90 days in arrears not restructured

Three-quarters of mortgages over three months in arrears at Ireland’s six main lenders have yet to be restructured, figures from the Department of Finance show.

Banks failing to tackle mortgage arrears

Three in four home loans in long-term arrears are not being restructured by the banks despite lenders having a range of mortgage solutions to offer borrowers.

The data shows 62,210 of these long-term arrears mortgages were not restructured, while solutions were agreed for just 20,424 loans.

It is very unlikely that all loans in arrears will, or even can, be restructured.  It is possible that a borrower who has historical arrears may now be back “on track” and hence a restructuring is unnecessary.  This highlights another shortcoming with the arrears data – they do not tell us whether a borrower is accumulating arrears now.  A measure of the number of accounts in arrears now, or even better, a measure of the number that are currently covering the interest on their loans would be very very useful.

It is also the case that there are many loans where a restructuring is just not possible.  For example, Ulster Bank have said that 35% of their customers who are in 90 day arrears are either not engaging or are not paying anything towards their mortgage.  It is nearly impossible to put a restructure in place in such circumstances.

It is also worth noting that a further 50,672 mortgages which are not in arrears have also been restructured.  This could be because any arrears has been repaid or recapitalised or because the restructure stopped the mortgage falling into arrears in the first place.  In total 71,086 mortgages in the sample have been restructured which is a lot of activity.

Finally, the data in this release represents 90% of all mortgage accounts in Ireland.  In this sample, 82,824 out 699,674 accounts were in 90 day arrears at the end of August.  That is 11.7% of mortgage accounts in the sample.

The most recent data from the FR for the entire mortgage market is for the end of June.  At that time, 12.7% of all mortgage accounts were in arrears of 90 days or more.  That means the tenth of mortgages that are in institutions outside the MART process have an arrears rate of 21.5% – nearly double the rate of arrears of those in the DoF sample.  The lenders here include INBS, BoSI, Start, Danske and although they are smaller than the other lenders in proportionate terms their arrears rates are far worse.

1 comment:

  1. "This highlights another shortcoming with the arrears data – they do not tell us whether a borrower is accumulating arrears now. A measure of the number of accounts in arrears now, or even better, a measure of the number that are currently covering the interest on their loans would be very very useful"

    Hi Seamus,

    It seems to me that our airwaves are filled with debates that lack anything like proper data.

    One of the great things about your blog is the level of data you extract from various government sources, however I think even then it is fair to say that these data often lead to even more questions and may have non-obvious inclusions, omissions & definitions that can actually mislead debate; Government accounts for one spring to mind.

    So how much do you think that the release of actual raw data would help analysts and public policy generally? What I mean here is basically spreadsheets and database dumps from banks and government. Not just in things like mortgage arrears, but tax returns, welfare payments and everything basically? Suitably anonymised (which may be more difficult that you'd think) would these data not be an antidote to spin, a boon to independent analysts and thereby to public debate and policy in general?