Features Australia


28 January 2017

9:00 AM

28 January 2017

9:00 AM

Centrelink – as everyone in Australia knows – shat the bed last month, leaving the government (and particularly Christian Porter) wrapped in a giant omelette.

Even outlets that traditionally take a hard line against ‘dole bludgers’ – like A Current Affair – recognised the stupendous scale of the cockup.

What was less well publicised is the extent to which hubristic faith in data matching and technology has become a substitute for good governance in both the public and private sector. Not only is Centrelink’s record keeping and data matching a mess, so is Telstra’s, and so is pretty much everything else we’ve sought to automate as we drown in so-called technical progress (and spam emails).

One friend of mine has taken to calling Telstra ‘the Centrelink of the private sector’ after a trawl through old bills, three hours on the phone, and travels through multiple telephone trees finally saw him credited 91 per cent of an unpaid, outrageously high bill.

He had to do this from a mate’s place as Telstra had cut off his phone and Internet.

That this problem is pervasive throughout organisations large and small, public and private, charitable and for-profit, gives the lie to the common classical liberal argument that government can’t be improved, and Centrelink’s snafu is the best we can hope for.

Australian governments are capable of efficient, if somewhat slow, service delivery. Indeed, it is characteristic of developed Westminster countries that governments are generally well run. This, after all, forms a goodly part of ‘governance’: yes, Virginia, there is more to it than verbal pyrotechnics in parliament. Time was when Australia was the envy of much of the developed world for its clean and well-run elections, its orderly and trustworthy census, the professionalism (if not friendliness) of its tax office, and the ease of establishing a company here.

No more. In the last five years, the AEC had to re-run the Western Australian Senate election, while ASIC blocked 250,000 websites due to a basic IP address misunderstanding (if one blocks one site’s IP address, other sites using the same IP address will also be blocked), the ATO’s data servers crashed, then we had the Census clusterf–k. And now Centrelink.

Small wonder suggestions Australia may move towards conducting its elections electronically were quietly shelved.

All save one of these cockups (the election re-run) came about thanks to technical failures – outsourcing gone wrong (ABS, ATO) or internal incompetence (Centrelink, ASIC). All were accompanied by blithe assurances that everything was fine and going to plan. At time of writing, Alan Tudge is still telling anyone who will listen that Centrelink’s debt recovery program is operating normally and not issuing fake debts.

It’s important to appreciate how broken is Centrelink’s new automated debt recovery regime.

When Centrelink’s initial letter invites people to confirm their income on MyGov, the in-house software – as its default setting – averages their income out fortnightly over the tax year.

This makes it look like people have earned money in fortnights when they’ve claimed benefits – particularly Newstart and Youth Allowance, although people claiming other benefits have also been stung.

It’s tempting and easy to agree to Centrelink’s figure if it tallies with the figure on one’s PAYG Summary (the old ‘Group Certificate’). I have no doubt – because it’s common to keep PAYG summaries for years – many people saw the figures matched and intuitively thought ‘oh, good, everything’s squared away’.

Next thing, those same people are getting texts from Dun & Bradstreet (‘rent-seeking bastards’ according to another friend) or having their Family Tax Benefit reduced from $170.66 to $1.00 in the name of debt repayment (‘kicked in the teeth for getting a job’, says a third friend).

Case studies (if you’re wondering where I sourced my Centrelink and Telstra stories) are easy to find. Thanks in large part to the Howard government, welfare provision has crept up the income scale, drawing more people into contact with government bodies – like Centrelink – willing to try their hand with dodgy data. I put out a call on Facebook (‘Anyone here been shafted by Centrelink? Answers on a postcard, please’), and got upwards of a dozen responses, with copious supporting documentation.

This means many of the people stung with fake debts are middle-class and have the capacity to hand over money they do not owe. About half the people who contacted me simply cut their losses (‘can’t sit on the phone for 3 hours’; ‘not taking time off to wait in a queue down at Centrelink’) and paid up.

However, this should not blind us to the fact Newstart – in the words of Tim Andrews, Executive Director of the Australian Taxpayers’ Alliance and a powerful advocate of fiscal restraint – ‘is a rounding error in the budget’. Many – perhaps most – of the people worst affected are being penalised for getting a job, or for moving between jobs. When it comes to perverse incentives, this strikes me as hard to beat.

From whence does this blind faith in data and technology come? Not from those who are technically knowledgeable. Data scientist Murray Neuzerling, tech journalist Asher Wolf, and IT consultant Justin Warren – among many others – have been at pains to point out that technology is a wonderful servant but a terrible master. For all the things computing power does, it still requires human oversight. Technology is not magic and there are many tasks to which it is unsuited.

Bureaucrats and CEOs make a grave mistake when they presume their organisation can ‘be like Uber’. They forget Uber became a start-up ‘unicorn’ because it was good. It took years of effort to iron out technical wrinkles and own up to its mistakes. Great tech is not made on the back of a fag packet.

We have to abandon the blind belief that we can ‘technologise’ everything. ‘Near enough for government work’ isn’t good enough if we’re going to leave crucial services in the hands of the machines.

Good governance comes first. After that, the rest follows.

Got something to add? Join the discussion and comment below.

You might disagree with half of it, but you’ll enjoy reading all of it. Try your first 10 weeks for just $10

Show comments