Mortgage Controls, the ECB and the Irish Housing market

Ireland’s monetary policy is set by the ECB and has had a very significant impact on household income and wealth in Ireland over recent years, as well as a profound effect on the housing market, particularly in relation to house prices and the ownership of the housing stock. Yet little attention is paid to it, in contrast to say Germany where it is heavily criticised and indeed the subject of legal challenge.  Similarly, the Central Bank introduced mortgage controls five years ago ,which again has had a material impact not only on housing but on credit growth, the distribution of wealth in Ireland and indeed on the political landscape, begetting policies designed to mitigate the consequences of these  monetary and macro prudential decisions.

Let’s start  with the controls. The average new mortgage for house purchase peaked in 2008 at €270,000 and then plunged alongside house prices before  bottoming in 2012 at €174,000. Since then it has risen steadily, reaching €233,000 last year, which when related to  rising incomes and the much lower cost of a mortgage indicates that affordability is much improved and in fact is still better than the long run average.

Mortgage controls are designed to limit  household leverage, imposing a LTI limit of 3.5 ( with some exceptions) and the Central Bank acknowledged in 2015 that  one might expect this to dampen house price inflation, credit growth and also negatively impact housing supply. We do not know how the market would have developed in the absence of such controls but the Central Bank  made a stab at answering  in their recent Financial Stability Review , estimating that prices in the period to the first quarter of 2019 would have been around 20% higher, or 4% per annum, with PDH lending substantially higher, by some 40%.The Bank does not show an estimate for housing supply but if prices had been higher completions would presumaly have been stronger, although what is also clear  is that the longer term relationship between house prices and supply has shifted since the crash, in that house building  has been much weaker in response to the actual price changes observed than experienced in the past.

If housing demand exceeds supply prices and/ or rents will increase, with that split being affected by, inter alia, the growth in income for  would be buyers, the cost and availability of credit and the type of buyer in the market. So if mortgage lending would have been higher in the absence of controls  then  some would-be buyers are forced to rent or live at home if that is an option. This has thrown up the odd situation  where the average rent nationally in 2019 was around €1200 per month, while the average monthly payment on a new  FTB 25-year mortgage  was  €1076 i.e. the rental payment could sustain a mortgage of  €254,000 instead of the actual FTB average last year of  €227,000. Landlords are therefore taking on higher credit risk than banks, and the average LTI for  FTBs  is actually only 3.1, which may be too low in an economy where the cost of housebuilding is high and where the  average rental payment  would service a mortgage with an LTI of 3.5 . It also  looks very conservative compared to the UK, where the LTI  cap is  4.5, with a 15% exception on a rolling twelve month basis rather than a calendar year.Rental growth in the UK has also been much weaker than in Ireland.

The type of buyer has also changed. The introduction of mortgage controls  coincided (?) with the ECB’s decision to buy bonds under QE, which alongside negative rates has pushed  Government bond yields into negative territory, including Irish debt out to 10 years. That renders Irish residential rentals yields ( which appear to be above 5%) unusually attractive and so we have had an influx of institutional buyers into the market, which was not a feature of previous cycles. Since the end of 2014  institutional buyers have purchased a quarter of new housing according to the CSO data, rising to 33% last year alone, which is then rented, with housebuilders also now more inclined to pre-sell developments to institutional buyers rather than risk waiting to sell to individuals.

QE is designed to boost investment in assets other than bonds and so will push up house prices, but  at the same time the Central Bank controls have constrained   access to mortgages for households. That not only has implications for owner occupation but also wealth: gross Irish household wealth rose to €947bn in the third quarter of 2019, of which €545bn was in the form of housing, or €412bn net of debt.Wealth in Ireland is therefore disproportionately held in property and so the combination of controls and QE has and will have  broader implications for wealth in Ireland  and its distribution over time.

As to the future, QE is open-ended at present and the market is not priced for a return to positive ECB  rates for years so the yield on bonds is unlikely to rise sharply, thereby maintaining the demand for rental yield. Similarly the Central Bank appears happy with the mortgage controls as they are, albeit showing some concern about the profitability of Irish banks (too low, that is), and if change occurs it may be to move towards a debt service metric- the impact of a fixed  LTI limit on mortgage payments  when rates are 3% would be be very different if rates were 5%.

A demand shock could change everything ( rents fell by over 20% from 2008 to 2010) as indeed could a supply shock. That might be positive (an upside surprise in terms of house completions) but also negative – an extension of rental controls or a rent freeze would reduce the value of the housing stock and it would be a very unusual economic development if that encouraged more completions.

Why has Irish house price inflation slowed?

Irish residential property price inflation has slowed appreciably this year. According to the CSO the annual change in the average residential property  in September was just 1.1%,  against 8.5% a year earlier. Prices in Dublin are now falling, by an annual 1.3%, and although still rising elsewhere in the country  the pace of appreciation has slowed to 3.6% from double digit gains a year earlier. Indeed in parts of the capital the falls are steep, with house prices in Dun Laoghaire /Rathdown 6.8% lower than a year ago, indicating that it is the more expensive areas that are seeing the largest falls.

What is driving the change? There are various approaches to modelling house prices and some straightforward metrics one can use to gauge whether residential property looks ‘cheap’ or ‘dear’, although there are data issues, including measuring the average price of a house- the CSO data is only available from 2010, for example, although the Department of the Environment has quarterly estimates going back much further, albeit solely based on mortgage lending.

One common metric is the house price to rent ratio, and on our model, using data back to 1990, the ratio is around the long run average and is actually falling currently, as rental growth in 2019 of 5.5% is outpacing house price inflation. House prices do look  more expensive relative to incomes, although interest rates are historically low with the result that the cost of an average new mortgage relative to income  is below the long run average i.e. affordability is supportive rather than excessive.

An alternative approach is to model house prices relative to ‘fundamentals’ , with the most common being mortgage rates,  disposable income, the housing stock and demographic changes. The Central Bank uses one such model (among a suite)  and  as pointed out in the Bank’s latest  Financial Stability Review it is currently pointing to prices  well below where  fundamentals suggest they should be. Moreover, the model  is predicting 12% price growth this year, so the recent slowdown is not readily explainable.

Our own fundamental model leads to a similar conclusion. Mortgage rates have edged lower over the past five years but the major positive  demand driver for house prices in the model is the very strong growth seen in  household disposable income , with both employment and wages rising at a strong pace, a trend still apparent. An increase in the supply of houses should act as a brake and that is indeed often cited as a factor behind the slowdown but we find this not to be the case. The housing stock per head of population  has the most significant impact and although completions are rising ,and may exceed 20,000 this year, that adds only 1% to the housing stock against  annual population growth of 1.3%. Indeed the housing stock per head has been falling since 2010, and so is acting to push prices up rather than dampen price growth.

So if such fundamental models imply prices should be higher there must be something else at work which is not captured in the regressions. Credit growth is one obvious factor, and we have seen  a change there  as the Central Bank has put limits on new mortgage lending since 2015. The Bank  believes that new PDH lending would be much higher in the absence of  such controls, as would house prices, although Irish banks are now  required to hold much more capital  so one might argue that lending would not have risen as quickly as in the past given the additional cost to the banks of  every new loan. The Central Bank actually voiced concern about the profitability of Irish banks in the Stability Review, which sits uneasily with the narrative commonly observed in many parts  of the media.

Another factor not captured in models is the type of buyer, and in Ireland’s case there has been a significant increase in ‘sale for rent’, with builders agreeing a block price for a development with an institutional  buyer attracted by the high yields on offer. The average price per unit sold is below what would have been achieved if each had been sold separately, acting to depress prices.

Sentiment and expectations  also play a role and not readily captured in a model.  Brexit and its potential negative impact on the Irish economy is an obvious factor here and consumer confidence surveys have certainly shown a significant fall this year. That uncertainty may be relieved to some extent if the UK actually  announces its withdrawal in January but then the issue switches to the type of trade relationship post- Brexit, so potentially prolonging the negative impact on sentiment.

If that were the case house prices could continue to ‘underperform’  the fundamentals  regardless of how the economy perfoms. Risks also abound in terms of employment and income while the current consensus view on housing supply may well be too optimistic, as softer house prices will dampen completions and hence continue to depress the housing stock per capita.

 

 

New supply data implies housebuilding consensus may be too optimistic.

Annual house price inflation accelerated  to 13% in April, the strongest pace for three years, bringing the increase from the cycle low to 76%, albeit still leaving prices nationally some 20% below the previous peak. Credit does not appear to have played a significant role in this recovery, at least to date ( net mortgage lending has only recently stopped falling and new lending appears to account for only half of transactions) and analysts have emphasisied more fundamental factors on the demand and supply side of the housing market.

In that model  housing demand has exceeded supply for some time now with the implication that prices will continue to rise ( absent a shock to employment or a steep rise in interest rates) until housing supply has picked up to a level which brings balance to the market. Supply, as in new housing completions, is picking up from historically low levels and in time is generally expected to approach and then meet estimates of annual housing demand. For example, the ESRI expect completions to exceed 23,000 this year and to reach 37,000 by 2020.

That expectation was based on completions data from ESB connections which put the annual figure at over 19,000 last year, although some thought that overstated the actual flow of new properties. Fortunately, the CSO has now started to publish completion figures on a quarterly basis, based on a range of sources , and they imply that completions may be much lower than generally expected over the next few years. The new figures, dating back to the first quarter of 2011, show completions of  57,000 over the past seven years, against an ESB total of over 90,000, a difference of 33,000. Last year’s total is now put at under 14,500, or 5,000 less than the ESB figure, and for this year the CSO figure  for q1 is 3,500. The latter may have been adversely affected by the weather but a full year figure of 17,000 or so seems much more likely than  anything approaching 25,000.

The implication is that the housing stock has been growing at a slower pace than previously thought  and that the stock  per head of population , a key variable in many demand/supply models, is still falling ( it started to decline in 2011) and will continue to decline for the next few years. Indeed, new population projections by the CSO highlight that demographic pressures will remain a feature of the Irish market; the population is projected to grow by over 300,000 by 2021 on a high net migration scenario or by 1.3% per annum, against a housing stock still growing by  well under 1%.

 

 

QE is fuelling Irish House Prices

Irish residential property prices have risen 60% since the lows of early 2013 but  this cycle is investor rather than credit driven. Gross mortgage lending for house purchase has picked but the average new loan has risen by just 28% over the past four years, implying a fall the average loan to value ratio, while data on transactions (recently revised up by the CSO) indicates that mortgage loans  still appear to be accounting for less than half of turnover in the market. New lending is also now constrained by the Central Bank’s mortgage controls.Moreover, net mortgage lending ( i.e new lending minus repayments) has been falling now for over seven years, although there are recent signs that it may finally be bottoming out.

Nothing here then to indicate that credit is playing a strong role in driving prices and it is curious that little attention has been paid to the impact of the ECB’s monetary policy  on the housing market, and, more specifically, its  non-standard measures including the asset purchase programme. The latter, QE, is designed to boost bond prices and hence lower yields so that ‘ investors may choose to take the funds they receive in exchange for assets sold to the ECB and invest them in other assets. By increasing demand for assets more broadly, this mechanism … pushes prices up and yields down, even for assets that are not directly targeted by the APP’.

QE is generally perceived as having a significant impact on equity markets and it would be odd if it did not therefore impact other  asset markets, including property, and we  can readily  see this at play in the Irish data on transactions. In 2011, investors (here defined as Buy to Let individuals  plus non-household buyers) accounted for 16% of residential transactions rising to 24% by 2012 and averaging a third of the market or more since 2014.

The yield on ‘risk-free’ assets , such as Government bonds, plays a big role in investment decisions and so the plunge in Irish Bond yields has been  a very significant backdrop for the Irish residential and indeed commercial property market : 10-year Irish yields peaked at double digit rates in mid 2011 but really started to fall sharply following Draghi’s ‘whatever it takes’ speech in 2012, and fell below 1% , where they still reside, following the commencement of QE in early 2015.

In contrast, the gross yield on residential property ( average rent/ house price) has not declined significantly in our rental model, and is still at 4.8%, having peaked at 5.4% in 2013. The rental yield fell to  a low of 2.75% during the last cycle, and is still well above the post EMU average (4.25%) and of course extraordinarily high relative to the ‘risk free’ rate available on Irish bonds, let alone Bunds.

The scale of investor interest in Irish property is therefore not surprising given the yield on offer and  is unlikely to disappear any time soon. Higher bond yields would make a difference, no doubt, and in that context the future of QE plays a part; the ECB will soon decide whether to scale back its asset purchases or indeed cease any additional buying. Yet it is likely to reinvest the proceeds of maturing bonds for a while at least, therefore maintaining the stock of QE, so absent an inflation shock bond yields may well stay low by historical standards. If so investor interest in Irish property will continue to be a big driver of the market.

Update on the Irish housing market

The Irish housing market has been characterised for some time now by excess demand, rising prices and  a record level in rents, although against a backdrop of contracting mortgage debt . Supply is increasing but  at a pace which is lagging the annual growth in demand, so it is difficult to see any change in the existing pattern, at least in the shorter term.

Indeed, house price inflation is now re-accelerating after a slowdown earlier in the year, according to the CSO’s new index. This is now based on all housing transactions, as opposed to those funded by mortgages alone, and showed a marked softening in the market over the winter months including a modest decline in prices in the three months to March. Momentum picked up again over the summer, however, with a 5.2% increase in prices in the three months to end-August, pulling the annual increase up to 7.2%. The earlier slowdown was most pronounced in Dublin and although prices have picked up again in the capital (the annual increase is now 4.5% from 2.6% in June) the re-acceleration has been clearly driven by developments in the rest of the country: prices ex  Dublin rose by 7.1% in the three months to end-August, taking the annual increase to 11.4%. That 3-month change is the strongest recorded on the index ( which goes back to 2005) and is reminiscent of the kind of price changes seen in the late 1990’s.It now seems likely that by December the annual increase in prices nationally  will be around 10%, which is stronger than many expected and compares with 4.6% in 2015.

Demand for housing would appear to be strengthening: net migration turned positive again  in the year to April, employment is rising by around 50,000 a year and wages are increasing again in the private sector, so helping to boost household income. In addition, mortgage rates are falling and our affordability model indicates that the cost of servicing a new mortgage relative to income is at levels last seen in the late 1990’s. New mortgage lending is indeed picking up, after a softer period post the introdution of the Central Bank’s macroprudential controls, but the increase is modest; some 17,300 loans for house purchase were drawn down in the first nine months of 2016, against 16,900 in the same period of 2015. For this year as a whole we expect a total of around 24,500 or 3% above the previous year. In value terms that equates to €4.8bn, and €5.3bn for total mortgage lending ( which includes top-ups and re-mortaging, with the latter rising rapidly in percentage terms, albeit from a very low base).

New lending is still being offset by debt repayment and this deleveraging has been evident now for six and a half years, although the most recent data does indicate that the pace of credit contraction is slowing. Another unusual feature of the market has been the preponderance of non-mortgage buyers , accounting for 50% or more of transactions. The third quarter data  currently  indicates that mortgage loans accounted for over 56% of transactions as recorded in the Property Price Register, perhaps indicating a slight change, although it is too early to say as the numbers on the Register are continually updated.

Residential rents have been growing at a steady 8%-10% annual pace for some three years now and the latest CSO  data, for September, shows little change, with an annual 9.6% increase  despite the Government’s rent controls.

What about supply, which is universally recognised as inadequate. Completions in the first eight months of the year amounted to just over 9,100, with the full-year figure likely to be around 14,500 or less than 2,000 above the previous year. Forward looking indicators do not signal any dramatic change, with  planning permissions for 6,200 units granted in the first 6 months of 2016.

Housing supply may well respond to higher prices in time but there is no quick fix to the current position of excess demand. In that context the announcement of a Help to Buy scheme in the recent Budget is hard to fathom, as it offers First Time Buyers a tax rebate towards their deposit. so presumably boosting demand further.The Finance Minister suggested that this would help to stimulate new builds  but it’s hard to argue that demand is the issue, rather than supply.

 

 

 

 

 

 

 

 

 

 

 

Irish mortgage controls having big impact on credit and Dublin prices

The Irish Central bank introduced controls on mortgage lending a year ago, including an 80% loan to  value limit on most owner-occupied properties , alongside a 3.5 loan to income ceiling. First Time buyers can borrow up to 90% loan to value for a property below €220k, which implied that the partial exemption would only be relevant outside the Capital; prices in Dublin probably average around €300k, against €175k across the rest of the country. The Bank’s own research suggested the controls would dampen credit growth and reduce housing supply, with a limited impact on prices. There is undoubtedly a case for such controls, particularly in an era of historically low  interest rates, but the timing appeared questionable given that credit has been contracting in Ireland since early 2010, with any new lending more than  offset by repayments.

The latest data, just published, indicates that the controls are having a pronounced affect on  new lending and are impacting house prices. Mortgage approvals for house purchase had been rising strongly (by an annual 40% in q1 2015 for example), albeit from a very low base, but that growth has stopped and approvals are now falling sharply; the annual decline in the final quarter was 20.3%, with December alone showing a 23.7% fall. There is not a consistent relationship between approvals and drawdowns but the trend in the former implies around 6,000 loans for  house purchase  in q4 against over 6,900 a year earlier.

One would expect the controls to bite harder in Dublin than elsewhere and the December data on residential property prices supports that view. Prices in the Capital , having risen by over 22% through 2014, fell in the first quarter of 2015 before regaining some momentum over the summer months and then fell again in the final quarter, by 0.7%, leaving the annual increase in December at just 2.6%. Prices ex-Dublin also fell marginally in the first quarter, implying an expectation  effect from the controls, but  picked up strong momentum in the latter months of the year. Indeed, the 5.8% rise in the three months to October was the strongest recorded by the CSO index, which starts in 2005, and other evidence shows that one has to go back to the late 1990’s for comparable gains. The pace of growth has slowed a little, with prices rising by  3.6% over the final three months of the year, leaving the annual increase in December at 10.2%, the same as in 2014. Some slowdown in Dublin prices was no doubt inevitable but the contrast between the Capital and elsewhere is striking, indicating that  would-be buyers in Dublin may be looking further afield.

The Central Bank has indicated that it will assess the impact of the controls in mid-2016 although the pace of contraction in new lending may prompt a speedier review, as one doubts it was anticipated.

Irish Mortgage Regulations impacting housing market

In late January the Irish Central Bank announced a set of macro-prudential controls on mortgage lending, Similar regulations have been introduced elsewhere, in line with the new orthodoxy in central banking, which  seeks measures to influence credit growth outside the traditional interest rate channel, particularly as rates are currently at historically low levels. The Irish version imposed a loan to value limit of 3.5 on Personal Dwelling Home (PDH) mortgages, but in the current Irish context the  second limit, on Loan to Value (LTV) was seen as a more binding constraint. A  maximum LTV of 80% is now in operation on PDH  mortgage loans, with first time buyers allowed 90% on properties up to €220k. Banks are allowed some discretion , but it is limited in that only 15% of loans can exceed these LTV ceilings.

Contrary to some commentary (and expectation), the controls were not seen as having a material impact on prices, and the Central Bank’s research showed that the  main effects would be on mortgage lending and the supply of new housing. Of course the controls would be pointless absent some effect on credit creation and in the Bank’s base case lending falls by 9% on the introduction of the new regulations and subsequently recovers some ground, although remaining below the benchmark case ( i.e. absent any controls) for over seven years.  In simple terms the new rules will require prospective buyers to save for longer, which also implies greater pressure on the rental market for any given level of housing demand.

Six months in, there is some evidence that the measures are having an impact across the housing market. Mortgage credit standards tightened appreciably in the first quarter and the latest Central Bank data shows that mortgage demand eased considerably in q2, from very buoyant levels over the past year.  That change is also evident in terms of mortgage approvals, with the annual increase slowing sharply in the three months to May, to 17%, from 41% in q1 and 56% in the final quarter of 2014 ( the latter  was probably affected by expectations ). Indeed, the annual rise in approvals in May alone was less than 8% and our own  mortgage models points to drawdowns for house purchase of 5.2k in q2, unchanged from the previous quarter.  New mortgage lending is still growing strongly on an annual basis but at a much slower pace.

Turnover in the housing market , which picked up very sharply in 2014, also appears to be slowing, based on data from the Property Price Register. Transactions amounted to 10.5k in the first quarter of 2015 and  also exceeded  10k in q2, but the annual rate of growth slowed to 13% from over 55%. The June figure was actually 7% down on the previous year and although late additions to the  Register are common the broad picture is unlikely to be seriously altered.

What about prices?  An unusual feature of the current upturn in residential values is the relatively high share of transactions (over 50%) driven by cash and so it would be surprising if the mortgage controls did have a very significant impact in that area. Dublin prices did fall in the first three months of the year, by 1.6%, but rose by 2% in q2, with a similar pattern evident in the rest of the country (a 2% rise following a 0.3% fall). The market has certainly cooled relative to the first half of 2014, but smaller price gains rather than outright falls appears to be the order of the day.

What about private sector rents?  Here, data from the CSO does point to an acceleration in what was already a buoyant market; rents rose by 1.7% in the three months to December but then picked up by 3% in the first quarter of 2015, followed by a 2.4% advance in q2. That means rents nationally are only 2% below the all-time highs recorded in 2008 and are therefore likely to surpass that figure by the final quarter of 2015.  As for housing supply it is too early to tell. although with only 2,600 completions in q1 the base figure is already very low by historical standards.

The central bank model predictions are therefore panning out in broad terms; mortgage demand has slowed, approvals have eased and transactions have  been affected , although  the impact on prices has not been dramatic.  In addition, the  upward trend in rents shows no signs of abating and that  perhaps  best illustrates  the real issue in the market- the shortage  of housing supply in the areas people want to live.

Bit early to blame Central Bank for house price fall

The CSO’s  residential property index   for February showed a fall in Dublin prices for the second month in a row, the 0.7% decline bringing the fall over three months to 2.4%. This still left the annual rise at over 21% but the market in the capital has clearly lost some momentum over recent months and some have claimed that the Central Bank’s new macro-prudential controls on mortgage lending are responsible. Prices excluding Dublin were flat in February but also fell on a three month basis, albeit by only 0.3%, so adding to the perception that there is a common factor at work across the country.

The evidence is not persuasive, however, at least not yet. The rules only came into operation in late January , for a start, and there does not appear to have been a significant shift in the recent pattern of mortgage approvals ahead of the decision. Mortgage approvals in the three months to January rose by an annual 55,5%, and as such not materially different from the 56.5% in the three months to December. Housing transactions in January were actually very strong, according to the Property Price Register(PPR), rising by an annual 68% . The available February data does show a marked deceleration in the pace of annual growth in transactions, to 35%, but that figure may be quite different when all the filings are included, which does take time.

The Central Bank’s own research (1) also suggests that the mortgage limits on Loan to Value and Loan to Income will have little impact on prices but a more significant  effect on mortgage lending and on the supply of housing, which they suggest will be some 2-3% lower per annum for a number of years ,resulting in a loss of some  2000  units  after 4 years relative  to an unchanged policy forecast. That  reduction in supply will put upward pressure on prices , so dampening any downward effect from tighter credit standards.

Any such simulation depends on the housing model used of course, and the ESRI (2) has just come out with some findings of its own. These also point to a significant effect from the new mortgage rules on house completions, with a supply fall of some 4%-5%, although they predict a larger effect than the Central Bank on prices, albeit  still a modest 4%-5%.

Another problem inherent in linking recent price trends in residential property to the Central Bank regulations is that not all housing is behaving the same way. Apartment prices nationally rose by 1.9% in February and by 2.5% on a three-month basis. Apartments in the capital also rose strongly on the month, by 2%, and by 1.8% over three months. The price series on apartments is extremely volatile but apartment prices in Dublin have now risen faster than houses over the past year (by 24.5% versus 21.1%).

Perhaps a better explanation for the most recent slowdown in house prices is simply that a market which appears to be primarily  driven by cash buyers is likely to lose momentum. That’s not to suggest that prices are likely to fall sharply but that annual house price growth in excess of 20% is unlikely to be repeated for long in the absence of excessive credit growth. New mortgage lending is picking up , and showing very strong percentage growth given the low base, but it is still accounting for less than 50% of housing transactions. Indeed, the latest PPR figures show transactions of over 15,600 in the final quarter of 2014, with the number of new mortgages drawn down for house purchase amounting to less than 7,000 , or 44% of the total.

(1) ‘Assessing the Impact of macro-prudential measures’ Central Bank of Ireland, Economic Letters , Vol. 2015, No.3

(2) ‘Quarterly Economic Commentary’, Spring 2015, ESRI