I have heard across the debt market that APRA has made two small but significant changes to how banks manage securitisations via APS 120. This may just kill Australian Lending FinTech.

End of Fintech

Two adjustments to the regulatory guidance have created new issues for those looking for warehouse financing or securitisation capability:

  1. APRA will limit any tranche retention in a securitisation to 20% notional of a senior tranche. In other words, a 20% senior exposure cap on what a bank can retain in a securitisation. There is a particular focus here because APRA doesn’t believe retained senior AAA and selling junior mezzanine tranches is a suitable risk transfer (ie. capital relief securitisation is not suitable).
  1. There is a loan warehouse growth cap of the lower of a bank’s investment home loan growth and the 10% APRA cap applied to investment home loans.

Major issue:

This will significantly impact all Lending FinTech at a later stage. I can’t see anyone scaling without either a banking licence or securitisation or probably both.

Funding Growth:

Lending FinTechs don’t have direct access to deposits, so if they want to scale and be cost effective, their only options are to tap into capital markets or other lenders’ balance sheets. And the most cost-effective way to do this is through bank warehousing facilities or loan securitisation. Further:

No “Unicorn” Lending or P2P FinTech exists (with the exception of RateSetter) without a publicly announced securitisation program.

See our table if you need further evidence:


Latest Valuation (USD) Securitisation Name


$1bn+ KABB


 $1.9bn+ CHAI
Funding Circle  $1bn+


Avant Credit

$2bn+ AVNT




OnDeck  $1.5bn


Lending Club  $7.2bn*


Zopa $760m+


Given lack of local Australian senior tranche or AAA buyers via capital markets (just 8% of super fund assets goes into non-bank fixed income, source: OECD), FinTechs would be forced towards the 4 Aussie majors. Overseas wholesale is just too much for a young local FinTech with a local product, plus currency swaps are an extra cost and complexity.

But now a growing FinTech cannot obtain funding from a major bank, as the capital relief type securitisation is now blocked. And don’t think you can try to run a book with 10 different senior tranche investors: new deals need to be especially “clubby”.

What about warehousing?

Warehousing is another option that will be constrained, although if there is not going to be a longer term securitisation market, the demand for warehousing might also fall. Warehousing is a stop-gap: helping non-banks get to scale before a securitisation. If this is now capped, then FinTech growth is now capped. I would also expect Australian banks to service existing clients with scale over high growth yet still small FinTech who might have longer-term payoff.

What Else? No deposits. No Collaboration.

Looping back, APRA has not showed any a desire to open up access to banking licences to FinTech (ie. direct deposit access) and the Federal Government has not got the appetite to provide support in the same way as the UK government did via buying loans or credit guarantees.

This, coupled with a low appetite for banks to engage with FinTech, has made life significantly harder.

What’s left?

In my opinion, not much. Small FinTechs with small warehousing facilities may exist but the ability to scale has now gone. Overseas FinTechs, who have already scaled, can come and then distribute securitisations given the bigger brand name and their local markets, notably the US, China and Europe.

Now APRA will have other reasons for its actions, so this isn’t a criticism of them. They need to manage Australia’s banking system. But this current approach appears to be pushing against where other global regulators and governments are going.

As it stands, the point of Lending FinTech was to bring technology and data into finance, which includes risk management and credit scoring processes for better borrower outcomes. This would push major banks harder, increase competition and, ultimately for Australians, create a financial services industry that we can export globally in the same way we consume a huge amount tech from Silicon Valley technology companies.

But it looks like APRA has now taken the Tech out of FinTech.



*We all know about Lending Club’s struggles over the last 2 years.

Read More

Beating the Robots


As a FinTech founder and mathematician, I should probably favour algorithms over human assessment. I often do. However I am getting tired of hearing about yet another algorithm that is able to predict, forecast and act on data to make financial decisions. Ultimately we’ve seen it before. Will the algorithms and robots win or does human analysis shine through for another business cycle?

This story starts with a cornerstone of basic AI and automated credit decisioning. Algorithms are built that take a lot of consumer data and perform analysis called logistic regression to develop a “good-fit” model. The observed data is then used to produce an equation that prescribes an expected behaviour. For lending, this is a probability of default forecast.

My first skepticism is due to my mathematics background. I loved engineering mathematics and fluid dynamics, when pure mathematics is used to derive equations that are fundamental to the development of aircraft, cars and all kinds of other high technology equipment we have today.

This mathematics saved time: a single equation developed through algebra meant we had ways to describe what would happen. The human mind equipped with paper was able to forecast without calculation.

As time went on, computational ability improved. In non-linear dynamics, we became interested into when and where systems went from calm to chaos. In effect, we wanted to know when our derived equations became unstable. Computational methods appeared that were able to map out these cases and build up smart eco-systems which then formed a more accurate forecast of reality than the human generated equations.

In the 1970s we started applying the initial human derived models without computational power to start predicting really annoying things like traffic jams. Over time, computers were introduced and we learnt that the computational part was much more influential that the initial models we derived: the human equations mattered less than the rapid ability for a computer to estimate all the potential ways the system would break down into a traffic jam – and then alleviate those risks. My last piece of research on this was in 2005 and computer power has move on exponentially.

Then in 2006 I did what all reasonably good engineers did: I went to work for Wall Street. Same equations and a real application of the mathematics (I failed to mention that the traffic jam research was mostly focused on driverless cars, something that will take another decade to become mainstream). Correlations, causations and logistic regressions where everywhere. In fact there was an attractive mix of human-built equations and computer generated ones. But the data source was bad: a long stretch without a recession and notably no significant house price corrections meant the instability was never tested. We know what happened next.

Moving forward to today, we see far less human developed equations. In any case, they have become far too complicated for many and I believe the variation of data in the world is significantly more important than the fine-tuning of the human equations, just like the traffic jam models.

But my concern now is about asking if these models are correct this time round? Can we trust the goodness of fit? Are sufficiently good stressed data sets used? Are people caring about the what-ifs? In my mind, the opportunity to beat the algorithms and robots is just as relevant as ever:

  • Australian banks view investor loans as less risk than owner-occupier loan. The data they use shows this. International evidence suggests this is not the case.
  • Data shows that SME lending is not worthwhile (and many banks have pulled back from this). But have they missed out on how new relationships can be formed or more suitable products can be developed.
  • Banks continue to adore their credit card businesses yet we all believe this rewards-based value capture will end at some point.

At the same time, we must also be aware that there will be human interference. Investor loans benefit from a tax break, SME lending will always be politically important and credit card use has been sticky.

The key to all this is not only capturing vast amounts of data but then finding ways​ that it can be effectively used to test your human or computer-generated equations but also how an artificially intelligent machine can adapt its equations when there will be known human interference, such as political and legislative changes.

This is how we can beat the existing robots.



Read More

Loan Denied: Tightening Loan Serviceability

Screen Shot 2017-04-05 at 11.45.59

We have written a few blogs on our view of residential investor mortgages and want to reinforce a few further points:

  1. Many banks view investor loans as lower risk than owner-occupier. This is wrong.
  2. Serviceability criteria will slow lending dramatically. ASIC is as much in-play as APRA.
  3. Serviceability is a yes/no result, meaning many investors will soon be blocked out from the market.

Investors loans have higher risk:

Most secured loan products for IRB banks follow the Basel PD, LGD formula to determine an unexpected loss value that is then a direct input into regulatory capital. The PD in this case is seen to be a through-the-cycle probability of default (TTC PD), which is then scaled up to a 1-in-1000 year stressed probability of default.

Australia has had benign lending environments in the last 2 decades, in effect no true credit cycle. So the first point to make is the TTC PD estimation is potentially a little low. In this data set, investor loans are showing a lower TTC PD than owner-occupiers*. One rationale banks believe is that rental properties have a replaceable occupant, so if one tenant loses his job, you can find another with a job.

However, the PD/LGD model is probably the wrong assessment for this asset class. Investment properties are revenue producing real estate, so you must wonder why residential properties get less stress than commercial property that gets put into Specialised Lending.

The PD scale-up is missing a systemic risk component that is not a spike in landlord defaults. It is an increase in rental property vacancy ratios.

Look at vacancy rates:

If an investment property is reliant on a tenant to cover the mortgage, usually interest-only, and the landlord doesn’t have a high capacity to cover the full mortgage payments themselves, then empty properties may lead to defaults.

So rather than looking at TTC PD, we should look at potential increases in rental property vacancy rates. Does a national movement from 2.5% to 12.5% seem appropriate for a 1-in-1000 risk event? We have seen unemployment rates move from 6% to 16%, so there is scope for at least a 10% increase in vacancy rates.

There is potential further stress in that a country-specific recession will have lower inbound migrants seeking work and expat communities might decline. Vacancy rates could increase further.

Added to this, rental prices will need to fall, due to excess supply of rental properties over demand. This is an unknown factor but will lead to further stress. As most investment property rental yields are below the mortgage interest rate, this is a dollar cost to the landlord’s income.

The owners of empty rental properties then face a dilemma. If they have a mortgage or higher loss to service, how will they pay for it? Relying on selling the property is not appropriate as 10% of other rental properties may be performing the same process at the same time.

Instinctively, 20% of investment properties becoming under severe stress feels about right. This is double the implied stress currently assumed on investor loans.

Serviceability is everything:

The above methodology is appropriate if the landlord is reliant on the tenant to cover the mortgage. If the landlord has a diversified source of other income, notably not just investment property, then you can assume they have the capacity to cover a single empty property.

The major issue with current serviceability calculations is they haircut the attainable rent by 20%. But as we described above, there could be an extra 10% of properties with no tenant and this is where the systemic risk is.

We expect ASIC and APRA to realise this risk at some point and tighten up the serviceability calculations so that investors reliant on filling a property will either not pass the serviceability assessment or they will be classified as Specialised Lending, which will have risk-weights 2 or 3 times higher and a mortgage rate several percentage points higher than owner-occupier home loans.

We look forward to these changes.


*Those that suggest this means there is lower risk in investor loans are incorrect. If expected losses do not appear, this is really a higher net income history rather than lower risk. Unexpected losses are the risk factor.


Read More

Why Is There Regulator Panic?

Screen Shot 2017-04-04 at 09.46.40

Panic is probably a little strong but APRA and ASIC are now appearing to take a much more vocal and active position on Australian housing risk. The RBA is also ready but whilst upward interest rate movement would reduce overheating Sydney and Melbourne markets, other aspects of the economy don’t need a rate hike.

Is there a bubble?

More people are saying there might be in Sydney and Melbourne, with another year of mid-teens house price increases. However, other fundamentals that should be a factor in house prices, namely wages, are at very low levels of growth.

There is a risk that this is credit driven price growth and we now believe there is a much higher risk of a large price correction.

Are interest rates to blame?

In part, yes. People base their ability to pay off a home loan by seeing a monthly mortgage payment. If they can pay that or cover it with rent from tenants, then it looks affordable. Interest rates are at an all time low, so unsurprisingly prices are at an all time high as affordability will always be stretched by those who want a house the most.

However, there are 3 things that make this a cyclical risk:

  1. Most of the debt is floating. Interest rates may and will increase due to the RBA or banks.
  2. A big proportion of the debt is interest only: the loan risk is staying there longer and borrowers are planning to not pay down debt in an orderly way (principal repayments might not be affordable in their eyes, be spent on other things).
  3. Rent isn’t covering the mortgage’s interest rates.

If we start to unpick these, we see that the regulators have allowed the proliferation of higher-risk borrowing.

Fixed is better than floating

There is no current long-term fixed rate mortgage market in Australia. The longest fixed period most banks offer is 5-years. This is not a long enough period for borrowers to pay down mortgage debt before the loan resets to a floating rate, meaning the borrower is still exposed to interest rate increases.

We need a 10-year fixed. We need longer. It’s a shame that the banking sector doesn’t understand the risk that not having this loan creates: banks believe they can just pass on interest rate increases to consumers without either higher defaults or a huge slump in consumer spending, which will lead to mortgage defaults as unemployment increases. This was one of the findings from APRA’s 2014 stress tests.

Interest Only is Subprime

The major problem with US subprime was that it relied on house-price appreciation to work. As soon as prices stopped increasing and teaser periods on Adjustable Rate Mortgages ended, there was a spike in defaults.

Interest-only investor loans are a similar risk. The borrowers have a negative rental yield as the mortgage rate is higher than the rental yield, so they only make money due to tax concessions and price appreciation. At 18.9% annual rate of increase, Sydney doesn’t have to wait too long before anyone living in their own home wouldn’t be able to afford their property if they were buying it today. Prices can’t increase forever

When prices stop rising, the trade unwinds or rental rates need to increase, leading to higher vacancy rates (people can’t afford the rent) or a diversion of money from consumer spending. Higher vacancy rates are what would lead to a huge correction in investor property portfolios. Higher rents, which cause consumer spending diversion, will lead to higher unemployment.

Serviceability is the key

We have no idea who is doing the serviceability assessment at the lenders but they clearly have an aggressive view on serviceability that is now a major driver for housing risks.

Providing floating rate debt with a buffer for 2 or 3% interest increases is simply not enough when we have seen interest rate cycles with 3, 4 or 5% overall increases in interest rates (look at the US Fed Funds Rate from 2004 to 2006 at +4.25%).

The key ingredient to a long-term fixed is that the serviceability risk is taken off the table: the interest rate can’t change. The risk with the loan can only reduce whereas floating or interest only loans can have higher risk as interest rates climb or the interest only period ends.

Will the RBA join the party:

In the long run, the RBA will have to assist in the unwinding of all-time low interest rates. However, it can’t increase interest rates with the all-time high debt load unless something happens:

  1. High inflation and high wage growth
  2. It actively tries to shift the floating rate debt to long-term fixed debt

And it is this second item that will eventually appear: the RBA will have to create a mechanism to control the longer part of the yield curve. However, the RBA, ASIC and APRA need to support the creation of a much broader wholesale market before this can happen and the banking system needs to understand the lower risk attributes of long-term fixed rate mortgages. We also need to hope that the non-bank sector doesn’t load up on the same risk and blow itself up in the process.

This also means regulators may have to actively support greater competition in the retail banking sector and ultimately reduce the sensitivity of consumers to interest rate increases.





Read More

Housing Deposit Call Options

One of the main features about US mortgages leading up to the GFC and subprime crisis was that borrowers were able to hand back the keys of their property. As prices fell and negative equity appeared, to many it was seen as better to default on your mortgage than pay off a debt that was more than the house’s market value.

Arguing the merits and risks of a “hand-back-the-keys” model is tough. Whilst it looks like a riskier lending option, it allows borrowers to default and move on to look for work elsewhere, in a severe recession without being locked into a debt they cannot afford. In any case, liquidity and commercial real estate is what kills banks.

In the US, the 100% or even 120% mortgage existed. A 100% mortgage essentially offered the home buyer the property in exchange for debt. As the borrower could walk away, this was then latterly described as giving borrowers a “free call option” on property prices. The 120% mortgage took this a little further as the property purchase had costs: taxes, decoration & refurbishment and advertising for a tenant (in some cases). You could describe the 120% mortgage as somewhere between a free call option on property prices or being paid for a free call option (you could argue being paid was a reflection that the property was overpriced).

And ultimately everyone loved a free call option and this led to large price increases and financial instability as people handed back the keys and homes to the the banks.

The question that needs to be asked: are free call options available today and is the “free” the problem?

What is a Free Call Option?

Firstly, there needs to be a definition: What is a true free call option vs. what someone perceives as a free call option?

Given the negative gearing and capital gains tax incentives, the shortage of land and appropriate new housing supply, being a property investor to many is seen as having a free call option on property:

  1. rent roughly covers the cost of mortgage (particularly after tax deductions)
  2. the downside risk is mostly not a factor

So investors are so positive that they believe they have the right to accept the gains in house prices but don’t face any cost. This is a natural part of credit and business cycles and the last quarter of a century has helped remove any perception of risk. Regular recessions and keeping price appreciation in check are good mechanisms here, but micro management of private asset prices is not seen as desirable (though I believe that price appreciation should form part of inflation measurement!).

In the case of investors, they are simply owning the asset and don’t have the ability to pass the asset back to the lender without consequence. So it isn’t a financial derivative.

Free Call Options:

This is where is gets technical, there is no free option as there are no 100% loans now that loopholes with personal loans have closed. What we do have is a potential paid call option in the off-the-plan market that may have the same risk profile. This is also where it gets mathematical.


  • S = Current Property Price = $250k
  • Deposit = 10% = $25k (what was paid)
  • t = Completion & Settlement date = 2 years
  • r = Risk Free Rate (Bank deposit Rate) = 3%
  • K = Strike Price = $225k

call price

Making various assumptions we obtain a call option price of between $25k and $40k, which is either equal to the deposit (very low volatility and setting yield = risk free rate) or that the buyer has bought an option worth $40k for $25k.

This isn’t a disaster and there is likelihood that the development profits easily cover the lost value in selling the option. However, options are more than value: where does the physical asset go?

Delivery of Asset

The key piece to all this is that in option trading you can end up being exposed to the underlying physical asset (taking delivery). This occurs if you cannot find a buyer to take the market price.

In the apartment market, the scenario where the would-be buyers don’t use their option is when the current market price falls below the strike price (a 10% price decline to below $225k in our example). At this point, the developer/seller needs to find another buyer. However, this may not occur as many would-be buyers may take the same option strategy and compound the price declines. The developer is then left holding the underlying properties, is unable to repay its bank loans (for the development) and the banks themselves might need to claim the collateral (the apartments).

If the banks themselves had short-term borrowings, then the collateral might also be a problem for them (liquidity squeeze).

So whilst we do see developers and banks (indirectly) selling call options at below market values, it is the potential delivery of the asset that causes the risk.

And this has more similarities with the US mortgage bust in 2006-2009 than we might think.

Read More

Gypsy Banking: Future of Finance

Are “Gypsy Banks” the future of Australian finance?

For 5 minutes, let’s forget about our banking oligopoly and look at another Australian industry which could also be considered oligopolistic.

The Beer Industry

With CUB and Lion Nathan dominating the beer industry, we can be certain that it isn’t as competitive as it should be. However, consumer trends have pushed back against the mainstream and enabled microbrewing to blossom. A small part of this trend is called Gypsy Brewing.


What is “Gypsy Brewing”?

Gypsy Brewing is when a small group of people hire out the facilities of a larger brewer to make their own batch of beer, usually because the larger brewer is not running at capacity. This could be due to:

  1. Demand for the beer made by a specific piece of machinery is down.
  2. Maximum utilisation is impossible with seasonality, so pockets of downtime periodically emerge.

The Gypsy Brewers get access to infrastructure they could not afford at such an early stage, test their recipes, manufacturing skills and sales and distribution of the new brews. People drink their beer and the beer industry evolves, with happy taste buds in the consumer, major breweries with a new revenue stream, and smaller brewers getting access to equipment they can’t yet afford.

And this isn’t limited to beer. A friend of mine at Pacific Ice Creams was testing his IP protected process for making UHT Ice Cream, using a major ice cream manufacturer’s facilities. With this he also got access to their expertise for ensuring the food passes consumption laws and regulations.

Why can’t we create Gypsy Banks?

So why is this limited to beer and food when a financial services equivalent would provide a huge amount of value? If this approach works to improve customer choice and deliver to different niches in beer and food, the same can surely be done for banking.

Our current banks could then be described in another way: Infrastructure Banks. They have the licensing and most importantly deposit financing (the two biggest barriers to a new entrant), but they are not operating at capacity.

We know the banks are not good at manufacturing new financial products, and in recent years have been more focused on product rationalisation. Their Legal, Risk, Operations and Compliance teams have seasonality, so staff won’t always be at full capacity. Other staff in the bank, may not be completely engaged, but giving them new and varied work may help keep them satisfied. And finally, as market conditions change, the banks will be looking for new revenue streams outside the core business.

So what would a Gypsy Bank look like?

This needs to be about manufacturing new products and services, which is not where banks are going with their innovation teams (who are focused on improving digital infrastructure and the customer experience). That excess seasonality in product compliance needs to be linked up with Gypsy Banks (FinTech) who have a product vision and want to sell it themselves. The existing banks should focus on providing their infrastructure at a price, be it 10, 15, 20 or even 30% return on capital. This needs to be efficient and transparent. 

So a FinTech needs to be in a position to say to a bank, “I have a product with a recipe (costs, risks, structure etc.)”. Then work with an Infrastructure Bank to agree a timeframe for product development, ideally around the downtime in Legal, Risk and Compliance (this shouldn’t be a marketing or business unit decision).

The Infrastructure Bank then needs to:

  1. Be able to agree to commercial terms for product development/prototype and medium-term manufacturing
  2. Look at contingency options if the Gypsy Bank proves a hit: is this an acquisition, is this a joint venture, are their extra services to offer (for example FX, hedging, complementary products).

Does this work overseas?

We are seeing this evolve in many forms. Tide is an example of a white-label small business bank run by Barclays. Other white-label products are further options but they are usually paid for upfront by the would-be Gypsy provider. Technically, they are not new businesses and in many cases are add-ons, such as mortgage broker group’s white-label home loans that have no distinguishing features to existing products.

The goal is to improve financial products and services so they provide the best possible outcome for consumers, including lower risk. People want or need different flavours.


Why banks should support this kind of structure and the competition it will bring?

This one is simple: local Australian FinTech are not a threat. FinTech globally are not a major threat. The major threat for our banks is coming from Google, Apple and Amazon from the US and the unknown quantity emerging from China by the likes of Tencent and Alibaba. After all, CUB and Lion Nathan aren’t worried about the threat of smaller craft brewers. They’re worried about other global giants or non-beer options coming to Australia to eat their lunch.

If we can then create a competitive market in Gypsy Banks and Infrastructure Banks, Australian consumers might be able to sample financial services that don’t leave a bitter taste in their mouth. And our banks will have access to a wider range of tools to compete with new entrants locally and offer to overseas markets.

I’ll drink to that.

Read More

Australian Investor Loans: How Much More Will They Cost?

rate delta

You don’t need to look far to see that property investor loan interest rates have increased significantly above owner-occupier home loans in the last year. Investor loans currently have interest rates 0.3% higher than owner-occupier rates.

Before APRA took initial action to limit year-on-year investor loan growth in 2014, investors and owner-occupiers were paying the same interest rate and with some bank preference to even charge investors less. This is at odds with overseas, where buy-to-let mortgage (investors loan) interest rates are approximately 1% higher per annum.

Part of this Australian market attribute is driven by the perception that investor loans have less risk: if the renter loses their job then you can replace them with someone else. If an owner-occupier loses their job, they might default on the loan. No recessions helps reinforce this.

Wrong Assessment

From my perspective, this is the wrong assessment and overseas banks clearly agree with me – hence the higher interest rate for buy-to-let mortgages.

For owner-occupiers, the 4 Major banks hold approximately 2.5% credit risk capital on home loans (APRA average). When offering home loans at 3.8% they make around a 30% return on regulatory capital.

The 2.5% credit risk capital is driven by the probability of default (PD) and loss given default (LGD). This falls into the ranges of 0.2% to 0.3% for PD and 10% to 25% for LGD. I expect the Major banks are on the aggressive end here.

For investor loans, there is a similar analysis with similar PD and LGD assessment. I also expect that the price increases are more of a supply/demand consideration given that the year-on-year limit is within sight – so there is a little else a bank can do except raise prices.

The Credit Risk Assessment is Potentially Wrong:

Key Question: Should investor loan credit risk be measured by Basel 2’s PD & LGD formula? Possibly not.

The main risk I see in these portfolios is the assumption that new renters will be available rather than the property investors’ ability to service the loan. So the 99% downside financial stress needs to be driven by 2 things:

  1. What will the renter vacancy rate increase to in a 99% downside scenario
  2. If there is rental income contraction, does that influence the likelihood of investor default.

These 2 items should be additive. Note: for credit risk capital we assume a fully diversified portfolio, so can use broad averages.

Renter Vacancy Rate:

As a broad estimate, the vacancy rate could increase by 10%. This will be driven by renters losing their jobs and/or downsizing and/or moving in with others for an interim during a deep recession. Unemployment rates could potentially rise by 10%, so I have mimicked this for the increase in rental vacancy rate.

I make no judgment on in investors themselves lose their jobs – but I would presume this would could to default. I will assume everyone who loses their job is a renter. If a property is not rented out for 6+ months, I assume the investor will default as property price declines will not bail them out (they will be in negative equity, on average). As unemployment rates will rise for more than 6 months, the vacancy and hence default rate will hit the 10% level.

Rental Income:

This one is harder as it is a sensitive supply & demand problem and the serviceability is a 2nd order impact. I will simply assume that there is not impact at this stage.

99% Stressed PD

Based on this, the 1-Year 99% downside Stressed PD (SPD) could be 5%-10%, depending on the increase in jobless rate, with a preference to selecting 10% and assuming a severe stress to the mining or financial sector.

The 10% SPD in this instance is also the stressed PD required for capital calculations. Using this in a loan capital & pricing model, I obtain an investor mortgage rate of 4.4%, which is 0.6% higher than owner-occupier.

Why is the UK 1% higher?

The calibration of loss-given-default might be important. Whilst broad market LGD could be assumed to be 20% (derived from a 35% decline in property prices plus 10% collection costs on an 75% LVR property), specific property investment stressed house prices could be considered to fall much further. If this is in the 50% to 70% range then the LGD is 40%+. If we then revisit our pricing, we can get a difference between owner-occupier and investors to match that 1%.

So What Happens Next?

Australia has a lack of recession-based data, so cannot really make an assessment of investor loan risk based on observed defaults and losses. The main concern is the divergence from countries and banks that have better data sets that have led to higher pricing for higher risk in investor loan markets. This higher pricing is also driven by stricter regulatory intervention which is also derived from a view on risk.

Given the regulatory hurdles in place, increasing risks for high rise apartments and the above analysis, we are forecasting that most Australian banks will have investor loans with interest rates 0.6% higher than equivalent owner-occupier. This is a further 0.3% increase on the current situation. Shame investors didn’t get a FlexiFix loan.

Watch this space!



Read More

Australia’s Major Problem: The 4 Major Banks Are Stuffed Full


One key attribute of the Australian financial system that people need to know about is the reliance on the 4 Major Banks to fund the economy. The lack of wholesale markets, driven by local insurance and local superannuation funds desire to invest sub-investment grade credit or equity portfolios (as opposed to investment grade and AAA/AA/A securitisations), has led to a reliance on the 4 Major Banks to fund the economy.

This is most easily seen in the following tables:

Total Assets $AUDm as Per Annual Reports
30/06/04 30/06/05 30/06/15 30/06/16
CBA 305,995 329,035 903,075 933,078
WP 237,036 254,355 812,156 839,202
NAB 411,309 419,588 955,052 777,622
ANZ 259,345 293,185 889,900 914,900
4 Major Total 1,213,685 1,296,163 3,560,183 3,464,802
GDP (past 4 Qs) 1,190,000 1,229,000 1,617,000 1,658,000
As % of GDP 101.99% 105.46% 220.17% 208.97%


Lending Liabilities $AUDm As Per Annual Reports
  30/06/04 30/06/05 30/06/15 30/06/16
CBA 205,946 235,849 674,466 700,547
WP 170,863 188,073 623,316 661,926
NAB 247,836 260,053 532,784 510,045
ANZ 204,962 230,952 570,200 575,900
4 Major Total 829,607 914,927 2,400,766 2,448,418
GDP (past 4 Qs) 1,190,000 1,229,000 1,617,000 1,658,000
As % of GDP 69.71% 74.44% 148.47% 147.67%

For the past decade, Major Bank balance sheets (both total assets and total loans) have grown, on average, at 10% CAGR (except NAB, which has been impacted by overseas businesses). These assets and loans are mostly domestic.

At the same time, GDP has grown at 3% CAGR in nominal terms, leading to approximately 7% higher bank balance sheet growth than GDP each year. The 4 Majors used to have total assets/loans at 100%/70% of GDP but now have 220%/148%.


At what point do the 4 Major Banks get full? How can the economy continue to grow at a strong level if the bank balance sheets can no longer grow credit above GDP growth as either deposit growth struggles to keep up or financial stability reduces given the high concentration risk of funding via such a narrow channel. Can this lead to a credit crunch? My view is that they may already be sufficiently full to prevent new lending at anything close to the rate experienced in the last decade (noting the below trend YoY growth in 2015/2016, albeit with NAB’s Clydsdale disposal making an impact).

Possible Mitigation

As a response, only a few alternative strategies are really available:

  1. Overseas Wholesale
  2. Local Wholesale
  3. Interest Rate Reduction
  4. RBA Funding (QE)

The 4 Major banks need to tactically grow overseas and local wholesale markets. Overseas distribution of securitisations, particularly the US market, has continued to be an available option but banks may need to consider how the pass fully desonsolidated deals into this market and what return requirements they are willing to take to facilitate this. Even if the 4 Major Banks don’t take the bulk of the spoils, the overall support this gives to the wider economy will lead to better outcomes for the 4 Majors. This may require supporting non-ADI Australian lenders, including FinTech, and identifying high-value opportunities in assisting these business. This may also lead to the development of broader local wholesale options.

Local wholesale potential is split between expanding the lending performed by insurance companies – which could be enhanced by regulatory changes – or by encouraging superannuation to take on a higher percentage of fixed income assets in their portfolios, particularly senior tranches of asset-backed securities. The obvious rationale: if super funds don’t support lending in Australia, there will be detrimental impacts to equity returns that make up the majority of their portfolios as net new lending is an input to economic growth.

Screen Shot 2016-11-16 at 13.54.22

An alternative approach will be for the RBA to buy financial assets. If a local securitisation market does expand, this could be supported by a form of QE. This would be inline with actions the US Fed took in buying agency RMBS or the ECB in corporate bonds.

The last option will be lower interest rates that may be linked with QE. Lower interest rates will support further credit growth and help banks manage a higher balance sheet level versus GDP. But note that interest rates are already low for an economy where the banks have mostly been well-run and capitalised.

Something has to happen

Without a system-oriented movement into addressing the runaway growth of the 4 Major Bank balance sheets, I would expect a major debt-driven problem to emerge in Australia within 5 years. This doesn’t imply a crisis at this point in time but does suggest an eventual fall in the rate that the economy can grow. If house price growth has been driven by credit expansion, this trend simply cannot continue without the financial system becoming top-heavy due to the 4 Major Banks’ balance sheets.

Read More

offset ripoff

Offset Home Loan: Is it worth it?

offset ripoff

One of the sacrifices we have to make when building a long-term fixed rate home loan is a reduced ability to add some extra features, such as an offset account to the loan. Whilst we don’t see this as a major issue (which we will show below) it was a sticking point with one bank we  talked to. They told us “the offset account is important in case borrowers want to buy a boat!”.


This drives us to offer value in other areas, with features like no break fees, freedom to repay a fixed loan anytime and a fixed rate for 10 years. These all form part of our key value: borrowers get to fix the interest rate for 10 years but can switch or pay off any amount when they choose. This has a huge amount of value.

The Offset

The history of the offset account is interesting and its use does make sense. Banks started offering offset accounts as a good way to combat non-bank lenders. An offset account, which is a deposit account and you need to be a bank or ADI to offer it, sits there and reduces your home loan exposure. If you have a salary being paid in, then the mortgage interest paid is reduced by the offset. More importantly, any savings you hold in deposit effectively attract the mortgage interest rate and are tax exempt (awesome).

For example (which we also use in our analysis later):

Home Loan Interest Rate = 4.0%

Best bank account savings rate = 2.5%

Tax Rate = 30%

If you didn’t have an offset, you would only save at 2.5% and be due to pay tax on those earnings, lowering the effective post-tax interest rate to 1.75% (2.5% * (1 – 30%)) for a 30% marginal tax rate payer.

Compare this to earning 4.0% with the offset account and looks like a good idea. But this is only a partial view.

Different Approaches

We need to be careful: the borrower could simply just pay off some of the mortgage with those savings and then refinance to release the savings at a later date.

Alternatively, the borrower could invest those savings into other assets and earn a higher return, albeit with some investment risk. Finally, banks may charge higher interest rates to get access to offset accounts.

What we want to know: how much is an offset worth and does it make sense to go for one?

To do this analysis, the impact of the Offset is measured across 2 fields: the Salary Impact and the Savings Impact.

Salary Impact

The first part to this is paying a salary into the offset account at the start of the month to reduce the loan balance. To make things easy, I will simply assume this salary is paid at the start of the month to give the largest possible impact. In practice, we would expect perhaps only 60% of the economic impact due to the loan balance reducing over the month – for example to pay off credit card and other bills.

We looked at 4 sets of households: low and high earners (in household terms) and each using an offset or not.

Household Household Income Mortgage Offset Account Estimated CPR Tax Rate
Low Earnings with Offset $60,000 300,000 YES 15% 25%
Low Earnings $60,000 300,000 NO 15% 25%
High Earnings with Offset $150,000 900,000 YES 15% 35%
High Earnings $150,000 900,000 NO 15% 35%

In addition, we also assume the non-offset household puts their salary into a savings account (at 2.5%) but pay tax on the interest earned.

To make a comparison, we reduce the interest rate of the non-offset household until the interest payments between the offset and non-offset households are essentially equal.

What we observe in these cases is the net benefit to the offset households is approximately 0.05% (5bps). If just savings methodology are applied, a household is 0.05% better off, each year, by using an offset account.

[To estimate this, consider $4000 paid in each month in salary. This equates to lowering the loan balance by $4000 for the life of the loan and saving 2.25% on that balance (4.0% minus 1.75%).  Assuming $300,000 loan balance, this is a $90 saving per year on $300,000 = 0.03%. We used a more complex model that considered inflation to assess the impact, hence the higher value of 0.05%.]

CPR made little impact over observable ranges.

If we assume only 60% of the economic impact, the annual economic impact is reduced to around 0.03%.

Savings Impact

The secondary impact is how savings influence the borrower and note that this is additional to the Salary Impact.

Assessment is harder as households could invest savings into non-cash assets. Cash in bank accounts is typically low earning as it is deemed to have very low risk whereas stocks and shares have higher risk but higher returns (and average to about 4-6% higher per year). Property has also increased significantly recently: should a borrower have savings or an investment property.

For our assessment, we will simply assume the borrower has savings in cash. For the offset account, there is a single approach as the mortgage balance reduces and the savings earn 4.0% without any tax implications (as 4% is the mortgage interest rate).

The non-offset borrow has 2 approaches:

  1. Pay off the mortgage and then redraw or refinance at a later stage as required
  2. Keep the savings in a separate account, get a lower interest rate (2.5%) and pay tax on the earnings

Within this, there are a few things to note. Paying off the mortgage is a good use of money if you have no desire to take additional investment risk in the short to medium term. If you do need to get access to the higher equity balance in your property, refinancing usually costs about $1500. You might get a free redraw access from some lenders (if you are not changing lender, the costs will be lower), so looking for redraw rather than offset is another option.

Keeping a separate savings account only makes sense if you want access to that money in the short term – notably within 2 years.

Combining the 2, if you know your approach, then we can find a maximum impact from not having an offset. If you need complete flexibility in the savings then there could be a far higher cost which we can then assess but do note: wanting full access and flexibility but not knowing how or when you will use it is possibly a poor approach – similar to leaving a large cash balance in your checking account.

Pay-off Mortgage and Re-Draw

The cost of this approach is based upon how often you re-draw/refinance and what the loan balance is. We assume each redraw $1500. We have set it out in years (how many years until the refinance/redraw is required):

Time (Years) 1 2 3 4 5 6
Cost ($) 1,500 1,500 1,500 1,500 1,500 1,500
Loan Balance ($) 500,000 500,000 500,000 500,000 500,000 500,000
Cost 0.30% 0.15% 0.10% 0.08% 0.06% 0.05%

We can then apply this across a range of values – with years along the top and the amount borrowed along the side.

$125,000 1.20% 0.60% 0.40% 0.30% 0.24% 0.20%
$250,000 0.60% 0.30% 0.20% 0.15% 0.12% 0.10%
$375,000 0.40% 0.20% 0.13% 0.10% 0.08% 0.07%
$500,000 0.30% 0.15% 0.10% 0.08% 0.06% 0.05%
$625,000 0.24% 0.12% 0.08% 0.06% 0.05% 0.04%
$750,000 0.20% 0.10% 0.07% 0.05% 0.04% 0.03%
$875,000 0.17% 0.09% 0.06% 0.04% 0.03% 0.03%
$1.0m 0.15% 0.08% 0.05% 0.04% 0.03% 0.03%
$1.125 0.13% 0.07% 0.04% 0.03% 0.03% 0.02%
$1.25m 0.12% 0.06% 0.04% 0.03% 0.02% 0.02%

I will revisit the results after Approach 2. Any borrower should take this approach as a maximum they should be willing to forgo for Approach 2: having flexibility costs money and this table shows you how much it should cost. It costs $1,500 to refinance and if the loan balance is higher or you plan to redraw very infrequently, simply wait until you need to redraw and pay the $1,500 – certainly if you save money elsewhere.

Keep Savings Separate, Pay the Taxes

Here, we assume you have savings as a percentage of the total loan balance, ranging from 1 to 20% and that borrowers will redraw every 2 years.

% of mortgage held as cash 1% 2% 4% 5% 10% 20%
Cash Balance ($) 5,000 10,000 20,000 25,000 50,000 100,000
Mortgage Rate 4.00% 4.00% 4.00% 4.00% 4.00% 4.00%
Savings Rate 2.50% 2.50% 2.50% 2.50% 2.50% 2.50%
Mortgage Cost ($) 200 400 800 1,000 2,000 4,000
Interest Earned ($) 125.00 250.00 500.00 625.00 1,250.00 2,500.00
Tax 30% 30% 30% 30% 30% 30%
After Tax Earnings 87.5 175 350 437.5 875 1750
Net Cost 113 225 450 563 1,125 2,250
Years 2 2 2 2 2 2
Total ($) 225 450 900 1,125 2,250 4,500
Cost per annum 0.05% 0.09% 0.18% 0.23% 0.45% 0.90%

These tables present a lot of data but it can be simplified into 3 sections:

If you have a very low savings balance (1-2%), using the offset has less value as you are not saving much money. We assume this is at best 0.1% per annum benefit in having an offset.

If you have reasonable savings, which comes to about 5% of your outstanding mortgage, the potential cost of this approach is 0.18% to 0.23% per annum (18bps to 23bps). From this point onwards, the decision is then a matter if the redraw will be made in Year 1 versus later. If you expect to redraw within a year, Approach 2 (savings & tax), otherwise Approach 1 (redraw/refinance). This caps the cost at about 0.15% (from Approach 1) as the 1-Year cost for Keep Savings in this second bracket is 0.1% and the maximum cost for Approach 1 is 0.15% if over 2 years (less if longer).

Generally, if one has a very large savings balance (10-20%) then there is limited point in borrowing. You are essentially giving the bank money by receiving 2.5% and paying 4%. Here, the best approach will be to use the redraw strategy (Approach 1) as the total cost is $1,500 and we expect the cost to be around 15bps as a maximum and reduce if you infrequently need access to the savings.

Drawing Conclusions

Offsets do add value to a borrower in that the tax benefit of reducing a liability is better than receiving interest and that interest us a lower rate than you pay.

By combining our assessment of Salary and Savings Impact, the offset puts a borrower in approximately 0.2% per annum better off: a borrower should be willing to pay up to this for the product.

There are several caveats around this:

If the savings balance is very low, the impact is reduced to about 10bps.

If the savings balance is proportionally higher, the borrower needs to consider what they are going to do with their savings: keeping large savings in cash deposits is not an optimal choice. If these savings are put into equities and earn 7% per annum, then the value of an offset account is reduced but the borrower is taking on a different financial risk that we cannot easily compare.

If you are a conservative borrower who doesn’t plan to redraw, the offset benefit may reduce to less than 10bps.

If you have other types of debt, you are better paying them off than reducing a mortgage, which typically has the lowest interest rate.

You may get access to free redraws in some home loans. Then benefit of an offset versus these is then reduced to 0.03%- 0.05% as we can use Approach 1 without facing the $1,500 cost.

Applying to a Family

Consider a family with household income of $150,000, 35% tax rate, $625,000 mortgage, $35,000 of savings and refinance frequency of 3-years.

An offset will save them:

Salary Impact: $2,264 or 0.05%

Savings Impact: Approach 1 is 0.08% (note: Approach 2 is 0.40%)

Total Impact is 0.13%, so they should only choose an Offset if the interest rate is less than 0.13% higher than a non-offset account.

What are banks charging?

The cost of an offset account ranges from 10bps to 40bps, so you need to consider if it is worthwhile versus going for the more simplified product. If a bank is charging more than 0.15% it is likely to be a financially negative decision to choose an Offset account (banks are known to gauge customers) as the bank is charging more than the likely benefit. We think most banks are charging 0.2% to 0.25% extra for offsets, so see them as being a pointless exercise unless combined with a split loan (we’ll save this for a separate blog).

Like many things, the borrowers view on how often they need to draw money is key and if the standard non-offset product allows the occasional redraw, this will help answer the question.

It may also be worth picking other loan products with no offset feature at all if they offer features that outweigh the offset’s value. The easiest example is a fixed rate product: if you think interest rates will increase, then a fixed offers potential value as those interest rates won’t go up during the fixed period.

Huffle’s FlexiFix home loan product goes beyond this. We offer the certainty of payments and the flexibility to pay off and refinance without any break costs. We believe this has substantially more value to a borrower than an offset account.









Read More

Mortgage Rate Forecasting, January 2017 (Pre-Trump)

With the clock winding down to the Inauguration of US President Elect, Donald Trump, a quick update on Australian mortgage rate forecasting.

In the last 2 months we have seen most Australian lenders increase their home loan rates across their fixed products (2, 3 and 5-year fixed rates) and well as increases to variable rate products. This is inline with our prior forecast in November. Investor loans have had a higher hit, in part due to higher systemic risk, the APRA 10% annual volume increase limit (banks using price to temper strong demand), and simply less sensitivity to rate increases (it’s not going to stop investors given high house price growth and the ability to negative gear away some of the higher increase).

Based upon maintaining a margin above where the cash rate is expected to be (2.25% premium), the lowest discounted variable rates are, on average, expected to reach 5.70% within 10 years. This is an average case forecast with variation possibly up to 2% higher.

Screen Shot 2017-01-18 at 09.44.45

As we have discusses before, now is a great time to borrow with fixed terms for as long as possible as rates are low, the RBA is expected to consider interest rate increases (1 in 3 chance of a rate hike this year) and incumbent banks are increasing interest rates even without RBA rate changes. The cheapest variable home loans have already increased about 0.2% in the last 2 months.

We believe the best way to mitigate the risk of higher interest rates is by opting for one of our 10-year FlexiFix mortgages. The interest rate is fixed for 10 years but borrowers can repay, refinance anytime without break fees after 3 years. If rates don’t go up, go down or you feel more comfortable in being able to manage interest rate rises, then you can switch without facing any finance related costs from our side.

This can be done with a straight mortgage or with a split loan (part FlexiFix, part variable).

Other Forecast:

We will look at what happens to expected rates after the Presidential Inauguration. Many experts expect to see a few interesting changes over the next few weeks.

To find out more and be the first to obtain the loan when we launch, visit www.huffle.com.au

Read More