How Do You Know What a Bank is Thinking? / by Adam Howard

Welcome to a New Year.  Best wishes for what, I am sure, will be a cracking year!! 

But then, we don’t know that for sure, do we?? I mean, we aren’t 100% sure 2017 will be a good one; healthy, safe, enjoyable, profitable…

And that’s it, right there.

Uncertainty. 

I have spoken in the past about the human race’s poor forecasting skills.  Our need to predict what will occur and our endless failures in attempting to do so.

If I wanted to be dramatic about it, I could say that organised religion is the result of our need to know what’s going to happen in the future.

Maybe 3 thousand years ago, or even more, a guy killed a rabbit just before his band of men had a fight with another band of men; a fight they won.

So, from that point on, every time they thought they might have a fight with another group of blokes they needed to kill a rabbit beforehand as that would ensure a win…..

And now we have the Catholic Church.

Anyway…

We are reasonable at forecasting simple events and contests, which technical speakers refer to as binary events; one with only 2 (maybe 3) possible outcomes.

Elections, sporting contests, stuff like that.

But even then, we get it wrong, and when we do the faceless crowd, the mob, is outraged.

Brexit, The Donald, The 2010 drawn AFL Grand Final….

Why did that happen?  How could we not have seen that coming?  What steps can we take to make sure that doesn’t happen again?

Humans crave predictability.  We are like children who love watching the same movie a dozen times.  Its safe, enjoyable and familiar, and we’ll even learn to enjoy the boredom in exchange for the stability.

A common mistake we make when we attempt to predict is confusing precision with accuracy.  Accuracy is defined as the degree to which a measurement conforms to the correct value, while precision is defined as a refinement in a measurement as represented by the digits given.

An easy example of what I am talking about is the US Presidential Election result.

Nate Silver is a bloke I have spoken about before.  A mathematician who made money as a professional poker player and became famous as a political forecaster in the US.  Link to his website is attached here:

http://fivethirtyeight.com/

The day before the election, Silver’s prediction was a 67% chance of a Clinton victory and 33% chance of a Trump victory.

The average person reads those odds and thinks “Clinton’s winning.  For sure.”  But those odds actually mean that if you held the election 3 times, Clinton wins twice and Trump wins once.  Pretty simple interpretation, but that’s it in a nutshell.

And guess what?  It was Trump’s 1 day in 3.

So maybe Silver was accurate in the odds he provided, which is what he was striving for:  accuracy.  He never mentioned or promised to precisely predict the outcome.

Another good example is the weather forecast.  It needs to be noted though that the weather is a slightly more dynamic system than an election.

I tend to visit the Bureau of Meteorology site as it gets updated regularly as new data is available and provides these probabilistic forecasts; something like, tomorrow there is a 70% chance of rain.

That then gets interpreted by the talking heads at the news stations who turn it into a forecast of showers tomorrow.

And that’s what the public believe; that it’s going to rain tomorrow.   100%.  No doubt about it.

Plans are made.  Picnics cancelled. 

And then it doesn’t rain and people are outraged.  How could the forecasters get it so wrong??  Idiots….

But they didn’t get it wrong.  The Bureau updates their forecast regularly and produces a forecast including outcome probabilities.  They don’t promise anything.  They leave that to the trusted TV weather reporter.

So, maybe it’s a problem with interpretation and not forecasting??

Maybe people are no good at looking at odds and understanding what they mean?  I think that is a good explanation as it explains why the gambling industry is so large and profitable.  It’s an industry made up of games and contests based on multiple possible and/or probable outcomes, with the market makers mostly understanding the probabilities and getting them about right.

And the players play, with the vast majority misunderstanding the odds.

But there are exceptions.  Leicester City winning the English Premier League last season??  That was a remarkable example of market failure.  Check out the story of the changing odds during the season on the link below:

The day before the election, Silver’s prediction was a 67% chance of a Clinton victory and 33% chance of a Trump victory.

The average person reads those odds and thinks “Clinton’s winning.  For sure.”  But those odds actually mean that if you held the election 3 times, Clinton wins twice and Trump wins once.  Pretty simple interpretation, but that’s it in a nutshell.

And guess what?  It was Trump’s 1 day in 3.

So maybe Silver was accurate in the odds he provided, which is what he was striving for:  accuracy.  He never mentioned or promised to precisely predict the outcome.

Another good example is the weather forecast.  It needs to be noted though that the weather is a slightly more dynamic system than an election.

I tend to visit the Bureau of Meteorology site as it gets updated regularly as new data is available and provides these probabilistic forecasts; something like, tomorrow there is a 70% chance of rain.

That then gets interpreted by the talking heads at the news stations who turn it into a forecast of showers tomorrow.

And that’s what the public believe; that it’s going to rain tomorrow.   100%.  No doubt about it.

Plans are made.  Picnics cancelled. 

And then it doesn’t rain and people are outraged.  How could the forecasters get it so wrong??  Idiots….

But they didn’t get it wrong.  The Bureau updates their forecast regularly and produces a forecast including outcome probabilities.  They don’t promise anything.  They leave that to the trusted TV weather reporter.

So, maybe it’s a problem with interpretation and not forecasting??

Maybe people are no good at looking at odds and understanding what they mean?  I think that is a good explanation as it explains why the gambling industry is so large and profitable.  It’s an industry made up of games and contests based on multiple possible and/or probable outcomes, with the market makers mostly understanding the probabilities and getting them about right.

And the players play, with the vast majority misunderstanding the odds.

But there are exceptions.  Leicester City winning the English Premier League last season??  That was a remarkable example of market failure.  Check out the story of the changing odds during the season on the link below:

The day before the election, Silver’s prediction was a 67% chance of a Clinton victory and 33% chance of a Trump victory.

The average person reads those odds and thinks “Clinton’s winning.  For sure.”  But those odds actually mean that if you held the election 3 times, Clinton wins twice and Trump wins once.  Pretty simple interpretation, but that’s it in a nutshell.

And guess what?  It was Trump’s 1 day in 3.

So maybe Silver was accurate in the odds he provided, which is what he was striving for:  accuracy.  He never mentioned or promised to precisely predict the outcome.

Another good example is the weather forecast.  It needs to be noted though that the weather is a slightly more dynamic system than an election.

I tend to visit the Bureau of Meteorology site as it gets updated regularly as new data is available and provides these probabilistic forecasts; something like, tomorrow there is a 70% chance of rain.

That then gets interpreted by the talking heads at the news stations who turn it into a forecast of showers tomorrow.

And that’s what the public believe; that it’s going to rain tomorrow.   100%.  No doubt about it.

Plans are made.  Picnics cancelled. 

And then it doesn’t rain and people are outraged.  How could the forecasters get it so wrong??  Idiots….

But they didn’t get it wrong.  The Bureau updates their forecast regularly and produces a forecast including outcome probabilities.  They don’t promise anything.  They leave that to the trusted TV weather reporter.

So, maybe it’s a problem with interpretation and not forecasting??

Maybe people are no good at looking at odds and understanding what they mean?  I think that is a good explanation as it explains why the gambling industry is so large and profitable.  It’s an industry made up of games and contests based on multiple possible and/or probable outcomes, with the market makers mostly understanding the probabilities and getting them about right.

And the players play, with the vast majority misunderstanding the odds.

But there are exceptions.  Leicester City winning the English Premier League last season??  That was a remarkable example of market failure.  Check out the story of the changing odds during the season on the link below:

http://www.skysports.com/football/news/11712/10261535/premier-league-201516-how-the-odds-changed-as-leicester-claimed-the-title

Leicester started the season at 5000-1 odds, which at the time seemed reasonable but, as with most extraordinary events when examined in hindsight, now seems ridiculous.

Interpreted, these odds meant that given 5,000 repetitions of the same event, Leicester City would finish on top just once…in a league of 20 teams that has only existed in its current form since 1991.

How can you possibly have 5000-1 odds when you have only had 24 instances on which to base those odds?  Sample size is too small…

And the odds were slow to change too, with Leicester City still at 100-1 a third of the way through the season while sitting in first place!!

Summary?  Extraordinary things happen and we underestimate the probability of them doing so.

Additional summary?  Less than extraordinary things happen all the time and we do a bad job of understanding the odds in a real world manner.

So, why bang on about this?  What’s my point this time around?

Well, I used to work for banks and so did my Dad.  Banks have a reputation for being boring, predictable.  And that’s the way people and the government like to think about them.

After all, that’s where your money is, right?  What YOU do with your money can be exciting and uncertain but when someone else is looking after it you want certainty.

And banks are safe…or are they?

Australia’s economy is relatively small by western standards, and has recently been ridiculously stable, having avoided the worst of the Asian economic crisis of 1997-98, the Dotcom crash of 2000-01 and the GFC of 2007-08.

So, you could say our economy is the exception, rather than the rule, and it also represents a fairly small sample set, having only 24 million people and GDP of $1.6 Trillion.

Might not be the best environment to test our “banks aren’t really that reliable and stable” hypothesis.

How about we look at the USA?  It’s the world’s largest economy with GDP of $17 Trillion and population of about $319 million.  Big sample set.

The USA has also experienced 3 recessions over the past 25 years, while Australia has experienced one, and over the past 150 years appears to average a recession about once every 4 or 5 years.

If you were talking Bell Curves, the USA’s economic cycles are probably a bit closer to the average over the past 25 years than Australia’s.

So, what level of bank failure has the USA experienced?

The Federal Deposit Insurance Corporation (FDIC) was established in the USA in 1933 following The Great Depression to insure customer cash deposits against the risk of bank failure, with the rationale being this would promote sound banking practices and confidence in the financial system (Yeah, right.  Has worked well, huh?).

The FDIC very kindly maintains a year-by-year list of the number of bank failures, which you can check out via the link attached below:

https://www.fdic.gov/bank/individual/failed/banklist.html

Here is a quick summary:

2016

- 23 bank failures

2015

- 15 bank failures

Period 2007 – 2016

- Total bank failures: 523

- Average failures per year: 52

The average number of bank failures is skewed due to the number of failures in 2009, 2010 and 2011 ( total of 389 failures)

- Average failures allowing for these three years: 19 failures.

That means that in the largest western liberal democracy on the planet about 20 banks fall over each year, and that’s in a market with about 5,100 banks.  That’s a failure rate of 0.4%; not far off the long term failure rate for home loan borrowers.

Now I know; it’s not the same country, economy, regulatory system, blah, blah, blah.  It’s just provided as an example.

Banks fail.  They do.  Just like businesses in any other industry fail.  And why?  Bad decision making by people.

Not because of short-selling.  Not because of poor economic conditions.  Not because of a black swan event.  Because executives, who are really just managers with fancy titles, make bad decisions.

And yes, something like this could happen in Australia. 

In 2008, at the height of the GFC, Bankwest was purchased by CommBank for $2 Billion; a fire sale by the then owner, Halifax Bank of Scotland (HBOS), who had attempted to make Bankwest the fifth pillar of the Australian banking industry.

Bankwest had grown at twice the industry growth rate for years and had done so by lending with NO MARGIN (at basically wholesale rates), expanding into the East Coast by chasing marginal, broker-introduced debt and not pricing risk appropriately.

Bankwest had been asking HBOS for more funds to meet operating costs for months, but when HBOS got crunched by the GFC (eventually being nationalised), Bankwest was left with no-one to help with funding operations, and started running out of cash….fast.

The Government brokered a rescue and CommBank got a screaming deal, buying Bankwest for much less than the net value of it’s assets.

This wasn’t a purchase; it was a bail out.  And it was a bail out because Bankwest came very close to failing.

No-one would have seen that coming.  No-one would have predicted that Bankwest, the home of Happy Banking, would have failed due to risky lending practices.

So, a quick summary is needed here. 

Banks fail, and generally it's because of poor decisions made by people. 

And, after holding deposits, the primary function of banks is to lend.  And due to capital adequacy requirements, a bank's loan book is normally about 10 times the size of the deposits it holds.

So, if a bank lends poorly and has to write off bad debts, this places those deposits at risk and the bank at risk of being insolvent.

This means that, by extension, banks can fail because of poor decisions made by people when deciding who is and isn’t credit worthy.  Who gets a loan, and who doesn’t.

And this is my question here regarding bank predictability.  Who does and who does not qualify for a loan?

The answer to this fairly straight forward question isn’t, actually, as straight forward as it should be.

I’ll provide a scenario which kind of makes my point, but before I do so, just remember the following:

The “3 Cs” of lending are:

Capacity: Do you have the financial capacity to make the necessary repayments with some comfort?

Collateral: Do you have acceptable collateral you can offer as security for the loan? And

Character:  Has your past behaviour shown you handle your financial obligations responsibly?

The final question you should probably ask yourself is “Would I lend this person my own money?”

My scenario involved an Australian husband and wife living and working overseas.  The wife worked in a highly paid role in the finance sector, while the husband had run his own successful financial services business for a few years. 

They had no debt and owned a property in the western suburbs of the Perth metro area which was fully paid for.  All details regarding their living expenses, working lives and income were fully disclosed and they had asked for a loan against the Perth property to engage in some potential investment activity, with their background showing they had done so successfully in the past.  

No issues in their lending past that would raise any concerns and the only thing that made this a little different was they wanted to borrow in the name of their family trust.

So, would you lend them your money?  All our 3 Cs appear to have been covered. I mean, I would have...

But, strangely, they couldn’t find a bank in Australia who would lend to them, at least not at a rate within 1% of the rate offered to other investors.

Unpredictable, huh?

All of a sudden our banks all decided that lending to self-employed expats wasn’t possible…at all.  Even if the proposal appeared to be strong, as the above did.

I pushed this as hard as I could, asking for a formal reason for the decision. 

The best I could get was a mid-level manager telling me he had discussed this with the local head of credit (the area where yes/no decisions are made) who had told him to “Tell him (me) to drop it.  We don’t like the application.” 

I admit it was hard to foresee this application would play out in this way, but then, I have a drawer FULL of other examples where banks behaved in unpredictable ways when deciding whether or not to support a loan application; both in favour of and against the borrower.

And so we come to it.  My point.  Don’t believe the ads that promote banks as predictable, benevolent sources of easy credit. 

Be a little skeptical and cynical. 

I am not suggesting we all become conspiracy theorists but perhaps that we adopt an evidence based approach to developing an opinion on banks and their motivations.

Maybe just remember an old saying that goes something like “A bank manager is a guy who will lend you an umbrella when it’s sunny, and ask for it back when it starts to rain…”