Cyclical Mortgage Controls might be better

The Irish Central Bank surprised many people, not least the Government I suspect,  by announcing changes to their macroprudential mortgage rules. The Bank had called for submissions on the controls , and received a good number, including one from this writer , but the rhetoric from Dame Street did not indicate a great appetite for change. In the event the Governor announced what was termed  ‘refinements to improve effectiveness’. Lenders have been granted  more discretion, in that 20% of total lending to non First Time Buyers (FTB) can be above the 80% Loan to Value (LTV)  ceiling , a change from the previous 15% discretionary figure, while FTB’s can now borrow up to 90% LTV , regardless of the price of the property (that limit was previously capped at €220,000). Moreover, up to 5% of lending to FTB’s can be above 90% LTV.

The Loan to Income (LTI) ceilings were unchanged, at 3.5,  and of course the LTV limits are just that; banks are not compelled to adjust their existing loan standards. Indeed, a recent Central Bank analysis of mortgage lending over the first half of 2016 showed that lenders were not fully utilising their available discretion to exceed either the LTI or LTV rules.

These changes also come in the wake of the Government’s Help to Buy scheme, which allowed FTB’s claim back income tax ( to a maximum of €20,000) to be used towards the deposit on a newly built home. One wonders if that subsidy would have been introduced had the new rules been in existence, as the combination of the two certainly provides a massive boost to FTB’s buying power in the property market. Take a couple with a combined income of €75,000,   borrowing €250,000 . Prior to the new rules they would have had to save €35,000 in order to buy a property worth €285,000,  perhaps taking four or five years. Now they could increase their loan to €256,500 (still inside the LTI limit)  and only have to find €8,500  as a deposit if the property was a new build ( a 10% deposit of €28,500 less the €20,000 from the Help to Buy scheme) implying it would take a far shorter time to amass that sum. Alternatively, all or part of their original savings could now be used to buy a more expensive house. In effect, purchasing power has been brought forward and leverage increased.

Credit appears to be driving only about half of current transactions in residential property but given the existing supply issues the boost to FTB’s buying power may well have some impact on price, although that is the mechanism which will eventually lead to a bigger supply response, albeit with a time lag.

The controls are designed to ‘enhance the resilience of both borrowers and the banking sector’ but are actually pro-cyclical; an 80% LTV still means the average size of a new mortgage will rise at the same pace as house prices ( for example, at a house price of €200,000 the mortgage would be €160,000  but rise to  €240,000 if house prices rose to €300,000). A cyclically adjusted LTV might be more appropriate, as put forward in our submission to the Central Bank. When house prices are low and credit growth weak, for example, the LTV limit might be 85% or 90%, but then decline as prices and credit growth pick up, to perhaps 75% or 70% at the top of the cycle. This would not eliminate cyclicality, but would dampen it, particularly if the LTI limits were also flexible over the cycle.

One obvious issue with this proposal is that the Central Bank would have to decide where we are in the credit and house price cycles. However, they currently have to do that anyway, at least in terms of credit; under new capital rules, some banks have to set aside additional capital ( counter-cyclical buffers) in good times in order to cushion losses in bad, and the Irish Central Bank sets this buffer every quarter, in part dependent on the current credit/GDP ratio relative to the long term trend. The buffer is currently set a  zero, reflecting the fact that credit is still contracting and the ratio to GDP is very low.

In addition, the Bank has a number of models that monitor residential prices relative to fundamentals. Deciding whether house prices are overvalued is not an exact science (  as it happens most models point to undervaluation , if anything, at this time) but if a number of indicators were to flash red the Bank could lower the LTV and LTI limits if a cyclically adjusted regime was in effect.

One final point. Research at the ECB supports the case that a debt service limit is more effective  in protecting borrowers and lenders than other macroprudential controls, as again outlined in our submission, although at the moment the absence of a credit register is a key impediment to implementation.


Mortgage arrears model points to further decline this year

Residential mortgage arrears in Ireland are extremely  high, both in absolute terms and relative to comparable housing markets.  At the peak of the cycle , 130k  mortgages were over 90 days in arrears , equivalent to  1 in 7 of outstanding mortgages owed to domestic lenders.  In the UK the figure peaked at a little over 1 in 100 and  in the first quarter of  2015 the total amounted to just 114k, in a market with 11.1 million mortgages. The good news is that the  Irish figure is now falling steadily, and our arrears model points to a further decline this year, in the absence of a significant shock to the economy.

Residential mortgages have been treated differently to other assets by the Irish  banks.  Real estate and commercial property loans were sold to Nama in 2010 for 43 cents in the euro, so crystallising a €42bn loss for the banks and opening up a capital hole subsequently largely filled by the Irish taxpayer.  Residential mortgages were not marked to market, in contrast, and arrears built up rapidly, reflecting , inter alia, societal pressure against large scale repossessions, the absence of foreclosure on any scale in modern Ireland,  some legal issues, political unease and a reluctance by the banks themselves in an environment of falling property prices and capital constraints.

Arrears on Private Dwelling Homes (PDH) are largely driven by three factors. The most important is unemployment, as the loss of a job and subsequent hit to income is one of the main reasons why mortgage payments cannot be met. The numbers unemployed in Ireland soared during the recession, from under 100k to a peak of 325k in late 2012, with the result that arrears  climbed rapidly. Interest rates matter too, although the impact is not as significant, and the decline in  mortgage rates since 2008 has had some offsetting impact on arrears. A third factor is house prices, perhaps surprisingly, but the relationship is clear in the data; the fall in residential values from 2007 to 2013 was a factor in pushing up arrears , with the scale of negative equity appearing to influence the decision on whether to continue to meet the monthly mortgage payment.

All three factors , with varying lags, help determine the level of PDH arrears in our model, which has performed reasonably well in tracking the 2013 peak and subsequent decline; PDH arrears in the first quarter of  2015 had fallen to 74k (9.7% of  the outstanding stock ) from a high of 99k (12.9%). House prices are now rising, so putting downward pressure on arrears , but the main driver of the fall is the improvement in the labour market and accompanying decline in the numbers out of work. As noted, these explanatory variables enter the equation with a lag so we can forecast arrears forward, given the current level of interest rates, unemployment and house prices, and that points to a figure around 50k by year-end, or well under 7% of the PDH mortgage stock. All econometric equations have a margin of error, of course, and debtor behaviour can change, particularly in response to  an economic shock or a perceived change in the attitude of lenders. The last few months has also seen a marked slowdown in the pace at which unemployment is falling, which if sustained will impact arrears into 2016.

There is less data  available on Buy to Let  (BTL) arrears and there seems to be other factors at work, making it difficult to derive a parsimonious model. Arrears in this market are proportionately much higher than private homes, although they  also appear to have peaked,  albeit a year after PDH, and are also now declining ; the q1  figure was  27k ( 19.7% of the total stock) from a  high of 32k (22.1%). The different drivers in BTL are also evident from the decision by lenders to send in rent receivers in order to recover mortgage payments, with the total rising to 6k in the first quarter.

The improvement in the economy and the recovery in the housing market have therefore resulted in a brighter picture on arrears, although these  factors have also prompted a change of tack on repossessions ( the sale of loan books, a return to bank  profitability and  a new  financial regulator in Frankfurt  have no doubt also played a part). The flow  of properties into repossession has certainly increased, rising to 557 in q1 from 354 a year earlier ( half the total is voluntary ) and that figure looks likely to rise, given the reported numbers before the courts.