It’s time to automate the ATM

Is an app-based self-driving car equipped with its own artificial intelligence-driven banking robot the future of the 50 year old ATM? Will the ‘automatic teller’ in the wall morph into an intelligent ‘bank-in-a-box’? Inter-connected with all manner of things and proffering all kinds of financial services?

Or does the ATM – described by a former chairman of the US central bank as the only decent banking innovation in five decades – even have a future?

This week the major Australian banks elected to wave the typical $A2 fee charged to customers using another bank’s ATMs.

"Will the ‘automatic teller’ in the wall morph into an intelligent ‘bank-in-a-box’?”- Andrew Cornell

Industry estimates put the fees lost to the industry at around $A300 to $A400 million a year and the move has been welcomed by politicians as a victory for consumers.

However, such ‘direct charges’ – made explicit when a withdrawal is about to made – were actually a policy decision by the Reserve Bank of Australia and its Payments System Board (PSB) in 2009 to drive competition in the ATM industry.

Customers were already being charged for using other banks’ ATMs, just not explicitly. The PSB argued a direct revenue stream would encourage more competition in the ATM sector and also encourage more deployment of ATMs in low traffic, higher-cost-to-service locations like rural and remote locations.

Opened up

The RBA’s latest ‘How Australians Pay’ survey noted easier access by competitors to the ATM sector and the explicit direct charging regime had opened up the sector.

“(Reforms) included that there are no unnecessary barriers to access for any potential new entrants and that the transparent direct charging model should remain in place for any ATM fees charged to cardholders,” the RBA said.

Indeed the most dramatic impact of the direct charging reforms was the increase in so-called “stand alone” ATMs – run by specialist ATM companies which did not offer banking services.

These are the machines often found inside convenience stores and other high traffic locations and their fees are typically around double the bank fee.

Direct charging represents the main revenue of such companies. Notably, shares in StarGroup Ltd, the only Australian listed company in the direct ATM business, slumped more than 20 per cent after the weekend announcement.

While StarGroup doesn’t have to change its fee structure, the share price response suggests investors believe there will be pressure to lower direct charges on all ATMs.

And there are existing and growing cost pressures on business using cash which have encouraged many to encourage non-cash payment.

Physical cash has no value until it is in an account so it is dead money. There are costs associated with security – transport must be secure; pilfering is rife.

Cash lubricates the cash economy and the Australian Federal Treasury has a ‘Black Economy Taskforce’ currently assessing the cost of cash to the economy.


The business model for ATM fleets is one thing. The other is consumer habits. As the RBA survey found, Australians are using cash less for transactions and electronic payments, such as cards and smart phones, more.

That is the case in almost all developed economies. Where ATM fleets are growing, it tends to be in emerging economies where unbanked populations are being banked – and cash use is growing.

According to the global ATM research specialists RBR in its Global ATM Market and Forecasts to 2021 report, the number of ATMs worldwide grew 5 per cent in 2015, reaching 3.2 million by year-end.

The majority of the growth came from China and India where RBR noted authorities are urging financial institutions to roll out ATMs to bring unbanked populations into the formal banking system. In the developed world, growth will be significantly lower.

RBR noted even in developed markets, ATMs and cash usage go hand-in-hand and cash remained resilient. The big question then is will cash remain resilient.

Just this week banks in Japan announced a new cyber ‘J Coin’ with the aim of reducing cash use. Japan is the most cash intensive major economy with hard currency usage at 70 per cent of transactions by value compared with around 30 per cent in others.

According the RBA survey “the trend decline in the share of consumer payments made in cash continued in 2016 – survey participants made 37 per cent of their payments in cash, compared with 47 per cent in 2013 and 69 per cent in 2007”.

“The decline in the use of cash relative to other payment methods since the 2013 survey mainly reflects consumers switching to contactless cards for lower-value payments. Nonetheless, cash was still frequently used for smaller transactions.”

In Australia, direct ATM charges have disconnected ATM usage from cash usage to a degree – but not in a way positive for ATM operators.

The PSB annual report notes “as transactional use of cash has declined, people are carrying less cash in their wallets and making fewer ATM withdrawals. Data reported to the Bank by financial institutions indicate that the number and value of ATM withdrawals declined by 7 per cent and 4 per cent, respectively, in 2016/17, faster rates of decline than in the previous few years”.

Meanwhile: “The average value of ATM withdrawals has steadily increased from around $180 in mid-2012 to around $215, possibly because cardholders are economising on their use of ATMs in order to avoid direct charges (emphasis mine).

“The rate of growth of ATM numbers has slowed in recent years, consistent with the decline in ATM use and associated pressures to rationalise networks; according to the Australian Payments Network (AusPayNet), there were around 32 000 ATMs in Australia at the end of March.”


Payments industry scrutineer Mike Ebstein of MWE consulting, who is preparing a major report on ATM usage, argues the introduction of direct charging has actually hastened the decline in ATM usage.

“We doubt it was coincidental that the peak annual ATM transaction numbers occurred immediately prior to the introduction of direct charging in March 2009 with the high point of 867.4 million withdrawals in the 12 months to January 2009 being 2.5 times the volume of 1995,” he told bluenotes.

“The ensuing decrease has seen numbers slip back to 622.6 million to the levels of 2002. The more overt nature of direct charges at ATMs no doubt contributed to the slowdown but we don't believe that is the basis for the ongoing decrease.”

Ebstein says the overall picture is one of a secular decline in cash usage.

“In particular, we note the drop in cash-outs at eftpos over the past four years,” he said. “Cash-outs at point of sale via the eftpos network account for more than a quarter of the volume of cash withdrawals and the decline in these no cost transactions has been primarily a result of the overall displacement of cash as contactless technology drives cards into lower value territory.”

Forecasting the death of the ATM – like the cheque – may be premature but even the most fantastic proposals for what it might evolve into over the next half century are inextricably linked to cash.

Mike Lee, chief executive of the ATM Industry Association, unsurprisingly sees a strong future.

“What is happening with cardless cash, iBeacons and app-based ATM automobiles is that the ATM is becoming more dynamic,” he wrote in an article on “The ATM at 100”.

“This transformation process in our industry is likely to accelerate when new artificial intelligence programs are incorporated into ATM software, expanding the range and types of services that will be available.”

It’s still about cash though: “cash, growing in terms of currency in circulation at rates significantly higher than average GDP rates, definitely isn't going into the museum in this generation or the next.”

“It has proved one of history's most endurable and popular technologies and has been the most dominant form of money ever. Given that cash is not "broken" and that there is no compelling business case to remove cash, I can safely assume it will still be needed in 2067.”

Maybe. More likely, particularly in developed countries, is a steady decline.

In banking history

So what of the ATM? Previously in banking history, when a technology has ceased to be a competitive differentiator and become merely a commodity, the industry has moved to a lower cost business model – often joint ventures.

This is what happened in the 90s with cheque-processing joint ventures which multiple banks joined; with early stage smart card platforms; with property conveyancing and myriad other ‘back-end’ functions.

Indeed the Australian banks had explored such a joint venture for ATMs where a dedicated, jointly owned utility would manage, service and stock ATMs (which would still be individually branded).

Removing half a billion dollars of revenue from the ATM business model is unlikely to see more machines rolled out. It is more likely to increase the pressure for such utilities.

Andrew Cornell is bluenotes manging editor

The views and opinions expressed in this communication are those of the author and may not necessarily state or reflect those of ANZ.

editor's picks

19 Sep 2017

Banking on climate change

Andrew Cornell | Past Managing Editor, bluenotes

There is another front - beyond politics and risk - for would-be financiers in the energy debate.

02 Aug 2017

Is fintech back (and did you know it had gone)?

Andrew Cornell | Past Managing Editor, bluenotes

For some, the fintech boom never subsided - but it’s clear its momentum and nature have shifted.