18 October 17

Australian banks still worthy of a place in most portfolios… despite what some commentators say

By Tim Farrelly – CEO of Farrelly’s, and asset allocation consultant to Australian Unity Personal Financial Services

Barring disasters, the banks should produce returns of the order of 10% per annum over the next decade. With a yield of 8% including franking credits, we need just 2% per annum growth to get us to a 10% per annum total return. Even if we get no growth in earnings, an 8% per annum return means that banks will be worth a place in most portfolios – barring disasters.

Disasters? What could possibly go wrong?

Anyone who follows the mainstream investment media will have no problem making some suggestions here. Ever increasing capital requirements, curbs on lending growth, new taxes, fines, Royal Commissions and other government interventions have been widely discussed. In addition, some outright disasters have been suggested, with a collapse in the residential property market the most common. And, of course, there is the possibility of an old fashioned, severe recession which inevitably would bring more pain for the banks.

Some of these scenarios are likely and should be factored into any forecast. Others may be unlikely but still are risks that we need to consider. Here, we want to put those risks in perspective particularly those that have been widely covered in the mainstream investment media and where we believe the impacts have been vastly overstated.

Increased regulatory and capital requirements

These are real and are happening right now and, accordingly, are in our base forecast. Most banks have around 10% of capital for each dollar of risk weighted assets – that should head towards 11% over time. This makes the banks safer but slightly less profitable. In addition, we have the bank levy which should slice around 2.5% off bank profits. Furthermore, we have threats of Royal Commissions, fines for bad behaviour, and so on. Collectively, we think these will reduce Earnings Per Share by about 10% over time. This slices just 1% per annum off returns over the next ten years. We include this impact in our forecast.

A slowdown in the growth of residential lending

We think this is highly likely and it is why we forecast future earnings growth at around 2% per annum. This is much lower than historical earnings growth and, in fact, this forecast is much lower than most other analysts’ forecasts. And still it gets us to a 10% per annum return.

A recession is likely in the next decade and will hurt the banks

Our forecasts assume that Australia will experience a recession in the next decade. We also predict that, when the recession comes, the market will know about it before we do – and so the chances of getting out early will be small. Hence, the key question is how bad a recession might be, both in terms of depth and also in terms of how well prepared the banks are for that recession.

The depth of a recession is often depends upon the health of the banks to that recession. The more extended the banks, the more they cut lending, the more they harass existing borrowers, and the more they drive the economy into the ground. When banks enter a recession in better shape, the recession is generally milder. We saw that during the GFC where the Australian downturn was much milder than in other parts of the world because, at least in part, the Australian banks entered the recession in reasonable shape.

A 2015 RBA study found that the key drivers of bank lending losses during recessions were: rapid credit growth; high levels of building construction activity; falling bank lending standards; and, rising interest rates.

Today, we have modest levels of lending growth, normal levels of commercial building construction, tightening lending standards and no sign of a central bank with any interest in raising interest rates. Of those four loss drivers, the only one flashing a warning light right now is the high level of residential construction activity. Even there, the banks are scaling back their involvement and watching their risks very closely. In short, the banks are in good shape generally and in much better shape than prior to the GFC. This suggests that any recession in the next decade should be relatively mild so long as these indicators remain strong. If they turn south, caution will be required.

Our forecast assumes that a mild recession will occur and will result in a one-off reduction in profits of around a third and take around 0.5% per annum off 10-year returns.

Even mild recessions will cause short-term volatility

But before we get too comfortable, we should not forget that during a recession, bank share prices will probably fall by 50% or more. But the fall is unlikely to be permanent.

While this may seem dramatic, we would say the same thing about every other sector of the share market. All equities are volatile. All can fall dramatically during recessions. The banks are no different. As long-term investors, we should worry predominantly about a permanent loss of capital.

And that is a possibility if the recession is severe. Accordingly, no matter how attractive the prospects of Australian banks, all the normal rules of diversification still apply.

Impact of a collapse in the housing market

Now, this is where things hot up. The market is divided on this issue. There are those who consider that a collapse in housing prices and as a result, the banks, is almost certain; there are those who  aren’t sure; and, there are those who are extremely sceptical that we will see a housing induced collapse in the banks at all.

Farrelly’s considers a collapse in housing prices as possible but unlikely:

  • We still seem to have a shortage of housing that not even the residential building boom is meeting;
  • Bank lending practices are being tightened but not sufficiently to cause an out-and-out collapse.


Nonetheless, it would be foolish to say that a collapse in housing prices couldn’t happen. Accordingly, we consider the impact of an extreme example – a 35% fall in the prices of houses nationwide and an accompanying recession that sees soaring unemployment and a 10% default rate amongst mortgagees.

Helpfully, the major banks produce detailed reports showing the Loan to Valuation Ratios (LVRs) of their mortgage lending books. This is all we need to do our own stress test.  Consider two loans, one has a LVR of 50% (in other words, $50 worth of loan for every $100 worth of house), while the other has an LVR of 90% ($90 worth of loan for every $100 worth of house.) Now assume that property prices fell by 35%.

Post the fall, the first loan now has $50 worth of loan for $65 worth of house, while the second has $90 of loan for every $65 worth of house. If the first borrower loses their job and can’t repay the loan, the bank has the option of putting the property on the market, recouping their $50 loan and sending whatever is left back to the unfortunate borrower.

The second borrower would be a problem for the bank. Here, a default potentially costs the bank a loss of $25 for every $90 of loan.

Now let’s assume that 10% of all mortgages default. The results for the major banks are shown in Figure 1 on the following page.

Figure 1: Bank stress test (35% downturn in property prices & 10% default rate)
Size of loan book ($ bill) 274 436 285 414
Loss as a % of loan book if 10% default -0.5% -0.6% -0.6% -0.4%
Loss In $mill -1,479 -2,660 -1,807 -1,482
Loss as a % of 2017 pre-tax profits -15% -19% -19% -13%
Pre-tax profits 2017 ($mill) 9,704 14,114 9,306 11,050
Source: Bank reports, farrelly’s analysis


That’s right. A perfect storm of a 35% fall in residential property prices and a 10% default rate would result in the banks’ profits falling by about 17% on average. While this is clearly not a great result, it falls a long way short of a disaster.

In a year or two, profits would rebound and normal business would resume. Farrelly’s calculations suggest that the whole episode would reduce 10-year average returns by around just 0.5% per annum.

Now, a much more likely scenario is that if residential property prices do fall that it will be more like a fall of around 20% (rather than 35%). This causes a one-off reduction in profits of closer to 4%. It’s a blip.

Residential property lending makes the banks safer, not riskier

The bottom line is this: residential property lending is actually an extremely profitable and safe activity for the banks. The fact that the Australian banks’ lending books are highly concentrated in home loan lending should be a source of comfort rather than concern. It’s the equivalent of having 70% of a portfolio invested in government bonds – the concentration, in this instance, makes the portfolio safer, not riskier.