The Fundamental Disconnect

You only have to look at the trends on housing credit growth and wage growth to see the problem. Using data from the ABS and RBA, we can see that in recent times credit to households for housing has been growing significantly faster than wage growth (note the two different scales) whereas in the 2000’s, before the GFC hit, wage growth was higher, and able to support such credit growth.  No wonder household debt is at a record high.

The other perspective is the cash rate, which has been cut to an all time low, and we see wage growth and rates trending in the same direction.

Whilst you can argue that lower rates means repayments are lower for many, the gearing effect of larger mortgages off the back of the home price boom, has created a major problem, with many households close to the edge at current low interest rates, and with low wage growth, no sign of this pressure relenting. Indeed, if rates rise (either officially or from market pressure) significantly more households will be stressed. Many will also struggle to pay off the capital.

These charts, together should have been a warning to regulators. The settings are wrong.

I think you can argue that we should be aiming for credit growth to match income growth, to stop the rot – but consider the impact on the banks (who rely on mortgage growth for profitability) home prices (a reduction in mortgage availability will force home prices lower) and housing affordability (credit rationing would lift the price of loans).

Even small adjustments might well create the conditions for a property market crash, with all the consequences that follow.

There is an old joke, when a driver stops to ask a local for directions, and the answer is “if I were you, I would not start from here”. The regulators have the same problem!

Latest Greater Perth Mortgage Stress Mapping

We now look across to the west, with mortgage stress modelling across WA, and mapping in Greater Perth. Read about our approach here.

The postcode with the higher number of households in mortgage stress in WA is Merriwa 6030, a suburb of South Western, Heartlands and  about 35 kms from Perth.  The average age here is 35 years and 40.6% of households have a mortgage. Some are in severe stress here.

Next is Samson, 6163, a suburb of Perth, Southern Suburbs, about 14 kms from the CBA and the federal electorate of Fremantle.  The average age here is 45 years and 36.8% of households have a mortgage.

Third is Carey Park, 6230, a suburb of South Western, South West, WA, and about 157 kms from Perth. 28.3% of households have a mortgage.

Next is Innaloo, 6018, a suburb of Perth, Other Western Suburbs and about 10 kms from the CBD. The average age here is 35 years and 32.8% of homes are mortgages. There are a number of households in serve stress here.

Meadow Springs follows on 6210, is a suburb of South Western, Heartlands, and about 61 kms from Perth. The average age is 37 years and 39.5% of households have a mortgage. Some are in serve stress here.

Note Wembley, 6014, a suburb of Perth about 5 kms from the CBD. The average age is 34, and 31.6% of households have a mortgage, but here there are the highest number in severe stress across Greater Perth.

Here is the geo-mapping for Greater Perth. The Blue areas are the post codes with the highest number of stressed households.

Latest Greater Melbourne Mortgage Stress Mapping

We turn our attention to VIC and Greater Melbourne with the latest mortgage stress mapping. Read about our approach here.

We are looking at owner occupied loans, with data to 1st March 2017.

First here are the top twenty post codes across Victoria with the largest numbers of households in mortgage stress.

Frankston, 3199 heads the list and is suburb of Melbourne on the Mornington Peninsula about 39 kms from the CBD. More than 30% of households have a mortgage here, and the average age is 38 years old.

Next is Berwick, 3806, a suburb of Melbourne, South East and about 41 kms from the CBD. More than 52% of households have a mortgage and the average age is 35 years. Berwick has a relatively higher proportion of severely stressed households (shown by the blue bar above). These households are in more immediate danger of potential default.

Third is Ballarat East, 3350, a suburb of South Western Victoria, Ballaratt, and about 99 kms from Melbourne. More than 30% of households here have a mortgage and the average age is 38 years.

Fourth is Rowville, 3178, a suburb of Melbourne, South East and about 26 kms from Melbourne. The average age is 36 years and 54% of households have a mortgage.

Fifth is Derrimut, 3030, a suburb of Melbourne, Geelong, and 18 kms from the CBD. The average age is 29 years and more than 70% of households have a mortgage. Note again the number of severely stressed households in the district.

Finally, note Carnegie, 3163, a suburb of Melbourne, East and about 11 kms from the CBD. 28% of households have a mortgage, and the average age is 35 years, but there are more severely stressed households here, than stressed. We think this is an indication of potential relatively higher risks.

Here is a stress geo-map for the areas around Greater Melbourne. The Blue areas shows the highest stress counts.

Next time we will look at Greater Perth.

Latest Greater Brisbane Mortgage Stress Mapping

We continue our series on mortgage stress mapping by looking at our results across QLD, with a focus on Brisbane. You can read about our approach to mortgage stress modelling here.

Harristown 4350 leads the list with more than 4,000 households caught. Harriston is about 109 kms from Brisbane near Toowoomba. Around and 28% have a mortgage. The average age is 37 years old.

Next is Manunda 4870, a suburb of Cairns, and about 1391 kms from Brisbane. The average age of the people in Manunda is 38 years of age and 21% of households have a mortgage.

Third is Geebung 4034, a suburb of Brisbane about 11 kms from the CBD. The average age of the people in Geebung is 37 years of age and 38% of households have a mortgage.

Here is the mortgage stress geo-map for the area around Brisbane.  The blue areas show those post codes with higher number of stressed households.

Next time we look at Victoria.

The Property Market, By The Numbers

In our latest video blog we walk though some of the most important numbers in the mortgage and property market, including the latest findings from our household surveys.

Some of the questions we answer are:

  • How big is the mortgage market?
  • How many borrowing households are there?
  • What is the average mortgage size?
  • How many households are excluded from the market?
  • What will happen if mortgage rates rise by 3%?
  • Where is mortgage stress worst?
  • How does the Bank of Mum and Dad in Australia compare with the UK?



Latest Greater Sydney Mortgage Stress Mapping

Following on from yesterday’s post, where we listed the post codes with the largest number of currently mortgage stressed households, we will now go into more detail across the main states. We commence this journey of pain by looking at NSW, where of course home prices have risen strongly, mortgages are large, and incomes static. Not a good recipe in a rising interest rate environment.

First we list out the top 20 post codes in stress across the state (by number of households in stress). The top three are Leumeah (2560), Chipping Norton (2170) and Bidwell (2770), all in the Sydney region. Next is South Tamworth (2340) and then Macmasters Beach on the Central Coast (2251).

Next, here is a geo-map of the Greater Sydney region with the counts of households shown. The blue areas are those with the highest counts.

We should make the point that mortgage stress measurement is getting at the current state of household finances, and is an indication the relative pressure on household budgets among households with owner occupied mortgages in the low wage growth economy. These in severe stress are nearer the edge in terms of mortgage repayments, but of course, given the significant capital growth in the region, most households will have sufficient equity to sell and repay the mortgage if need be. Thus, when we overlay our economic projections to derive a measure of default probability by end 2017, this is an indication of households missing a mortgage repayment (30-day default), not bank losses, which in our modelling remain extremely low.

The mortgage books in this region are bolstered by rising equity, Lenders Mortgage Insurance on higher LVR loans, many households paying ahead (though not those in stress) and current low interest rates. However, the scenario might change were rates to rise, home prices to slide and unemployment or underemployment were to rise.

We will look at Brisbane next time.


More than 1 million Australians face mortgage strife

From The AFR.

Exclusive analysis performed for AFR Weekend has revealed that more than a million Australian home owners will struggle with mortgage stress if interest rates were to rise just three percentage points.

Data from research house Digital Finance Analytics shows that close to one in three households from Victoria, Tasmania and Western Australia will experience mortgage stress ranging from mild to severe in the event of just three rises of 25 basis points. A rise of 300 basis points, back to more normal levels, would be much more severe.

The news follows warnings from the OECD that the nation has a one in five chance of entering recession and Australia’s $6.5 trillion housing market is running the risk of a hard landing.

Runaway house prices in Melbourne and Sydney have added to the risks facing the economy as rising levels of household debt make home owners and property speculators vulnerable to unexpected moves in interest rates.

The Reserve Bank of Australia’s official cash rate, on which mortgage rates are based, is as an “emergency low” 1.5 per cent. In practice that means mortgage repayment rates are between about 4 per cent and 5 per cent of the loan amount per year.

Digital Financial Analytics principal Martin North says that a shake out in the property market would not be restricted to lower income areas and would include households in the trophy suburbs of Bondi and Lane Cove in Sydney as well as homes in the leafy green streets of Toorak and Prahran in Melbourne.

“The common theme here is affluent households paying top dollar for apartments with big mortgages and the potential to be caught out by rising interest rates and flat or falling incomes. Even places like the lower north shore are being hit” he said.

Under the modelling performed by Digital Finance Analytics, there are around 650,000 households in Australia experiencing some form of mortgage stress. The numbers are consistent with a Roy Morgan report from September 2016 that showed one in five households were experiencing mortgage stress.

The longstanding measure for mortgage stress has been 30 per cent of household income.

Mild mortgage stress might see household cut back on childcare expenses, dipping into savings or reaching for the credit card in order to make payments. Severe mortgage stress indicates that the mortgage holder has missed a payment or payments and is already considering selling the property.

If rates were to rise 150 basis points the number of Australians in mortgage stress would rise to approximately 930,000 and if rates rose 300 basis points the number would rise to 1.1 million – or more than a third of all mortgages. A 300 basis point rise would take the cash rate to 4.5 per cent, still lower than the 4.75 per cent for most of 2011.

Professor Roger Wilkins of the Melbourne Institute at the University of Melbourne produces the Household Income and Labour Dynamics Survey, regarded as one of the best sources of information about housing affordability in Australia.

He says that while mortgage stress hasn’t materially increased in recent years that a sharp rise in interest rates would be destructive to household finances everywhere.

 “If the cash rate goes to 6 per cent then you would expect to see a lot people in strife. Particularly with wage growth and inflation at such low levels so that does increase vulnerability to rises to interest rates” Mr Wilkins said.

Mortgage Stress Grinds Higher

We have just rerun our mortgage stress models, incorporating data to 1st March 2017. Household budgets remain under pressure, thanks to flat incomes and rising living costs – and some lifts in mortgage rates. You can read more about our approach to measuring mortgage stress here. Our current analysis concentrates on owner occupied mortgages.

Overall, 21.78% of households are in some degree of mortgage stress. We look at mild stress, meaning they are managing to meet repayments, but are doing so by cutting back on other expenditure, putting more on credit cards, and seeking to refinance or restructure to reduce monthly payments.  19.08% of households fall into this group. An additional 2.7% of households are in severe stress, meaning they are likely to miss repayments, or are in default, or are looking to sell. We look at household cash flow, not a set percentage of income going on the mortgage.

Here are the postcodes across Australia with the highest levels of stress.

Harristown, QLD 4350, which was the highest count in December 2016, still is first, followed by Leumeah NSW 2560. Leumeah  is a suburb of Sydney, Macarthur/Camden, about 40 kms from the CBD.  The average age of the people in Leumeah is 35 years of age. Around 37% of households there have a home loan mortgage.

Third is Frankston VIC 3199, which is a suburb of Melbourne, Bayside, It is about 39 kms from Melbourne.  Frankston is in the federal electorate of Dunkley. The median/average age of the people in Frankston is 38 years of age. Around 30% of households here have a mortgage.

Fourth highest is Merriwa 6030, a suburb of South Western, Heartlands, WA. It is about 35 kms from Perth in the electorate of Pearce. 41% of homes here are mortgaged.  The average age of the people in Merriwa is 35 years of age.

Once again, remember interest rates are very low, and are expected to rise, so the OECD warning about the risks in the housing sector seem well placed.

Editors Note. We updated this post to reflect the total of mild and stress, when it first appeared, we sorted only on mild stress, which changed the results slightly – and we also added back the latest probability of default metrics as well.


More On Household Debt, From The ABC

ABC’s RN Breakfast‘s Business Reporter Michael Janda discussed household debt as part of his segment on Radio National Breakfast this morning, and was kind enough to mention our recent research on owner occupied and investment housing debt sensitivity.

There was a subsequent flurry on Twitter discussing the DFA research approach.

To be clear, our household modelling is based on a rolling 26,000 statistically robust omnibus survey, to which each month we add 2,000 new households and drop off the oldest set. We have data from more than 10 years of research and it feeds our programme of activity and is reflected in the DFA blog.

From a mortgage stress perspective, we run our modelling, based on our household profiles and segments, which looks at net cash flow (before tax) and we also sensitive the modelling based on potential future rate movements. We take account of their total financial position, including other debt demands, and costs of living.

You can read more about our modelling here.  If you want to read our mortgage stress work, this overview is a great place to start.

P.S. Our research is separate and distinct from other research in the housing affordability arena, including the international Demographia survey. Whilst some of the findings may align, the research is based on different underlying research sources.


So Just How Sensitive Are Property Investors To Rising Interest Rates Now?

Having looked at changes in investment loan supply, and the motivations of the rising number portfolio property investors, today we use updated data from our rolling household surveys to look at how property investors are positioned should mortgage rates rise. In fact, for many, rates have already been raised, thanks to lender repricing independent of any RBA cash rate move, some as much as 65 basis points. We think there is more to come, as loan supply gets tighter, international financial markets tighten and competitive dynamics allows for hikes to cover capital costs and to bolster margins.

To assess the sensitivity we model households ability to service mortgage debt, taking into account their other outgoings, and rental income.  We are not here looking at default risk, but net cash flow. How high would rates rise before they were under pressure? Where they also have owner occupied loans, or other debts, we take this into account in our assessment.

The first chart is a summary of all borrowing investor households. The horizontal scale is the amount by rates may rise, and for each scenario we make an assessment of the proportion of households impacted, on a cumulative basis. So as rates rise, more households would feel pain.

The summary shows that nationally around a quarter of households would struggle with a rate hike of up to 0.5%, and as rate rose higher, this rises to 50% with a 3% rate rise, though 40% could cope with even a rise of 7%.

So a varied picture. But it gets really interesting if you segment the analysis. Those who follow DFA will know we are a great believer in segmentation to gain insight!

A state by state analysis shows that households in NSW are most exposed to a small rate rise, with 36% estimated to be under pressure from a 0.5% rise (explained by large mortgages and static rental yields), compared with 2% in TAS.

Origination channel makes a difference, with those who used a mortgage broker or advisor (third party) more exposed compared with those who when direct to a lender. The pattern is consistent across the rate rise bands.  This could be explained by brokers knowing where to go to get the bigger loans, or the type of households going to brokers.

Households with interest only loans are 6% more exposed to a small rise, and this gap remains across our scenarios. No surprise, as interest only loans are more sensitive to rate movements. We have not here considered the tighter lending criteria now in play for interest only lending.

Our master segmentation reveals that it is Young Affluent and Young Growing Families who are most exposed, followed by Exclusive Professionals. Some of the more affluent are portfolio investors, so are more leveraged, despite larger incomes.

Finally, we can present the age band data, which shows that those aged 40-49 have the greatest exposure as rates rise, though young households are most sensitive to a small rise.  Note this does not reveal the relative number of investor across the age groups, just their relative sensitivity.

This all suggests that lenders need to get granular to understand the risks in the portfolio. Households need to have a strategy to prepare for rate rises and should not be fixated on the capital appreciation, at the expense of cash flow management, especially in a rising rate environment.