Revenge of the auditors

- First published in the Australian 17 November 2014.

IT’S now a little over a year since the Australian Electoral Commission discovered that 1, 370 ballot papers were missing from the recount of Western Australia’s Senate vote for the 2013 federal election. Worse, much worse, was that the closeness of the count made an expensive re-election necessary.

It was around this time that the AEC crawled into the doghouse. It has not yet been invited out.

Once the media has its teeth into a theme it is reluctant to let go, as we saw during the re-election in April when every rumoured AEC broken pencil, every disorderly queue was gleefully retold as further evidence of how hopeless this organisation is.

The Abbott government is going hard against the statutory authority, and the Labor opposition is not exactly rushing to its defence. AEC staff fronting the Joint Standing Committee into Electoral Matters tend to receive lots of finger-pointing and shouting, even more (from what I’ve seen) from ALP members than Coalition ones.

Everyone claims to be trying to “restore Australians’ faith” in the AEC, but this will not be possible as long as every mole hill is blown into a mountain.

All organisations make mistakes. Which newspaper doesn’t, for example, find the odd rogue apostrophe slipping through the cracks, or worse?

After the bungle was uncovered last year, the then commissioner Ed Killesteyn got former federal copper Mick Keelty to investigate. Keelty’s report, released in December, was quite damning.

It gets worse, because back in 2010 the Australian National Audit Office delivered the findings of an audit of the AEC which it now says have not been sufficiently acted on.

And this year the ANAO has conducted two “follow-up” audits of the AEC, the second of which was released early this month.

The approach of Acting Commissioner Tom Rogers seems to be to cop it sweet, accept all criticism, not argue, put hand on heart and assure JSCEM this has been a wake-up call, a “period of reflection” for the organisation. We know we’ve let everyone down and we’re determined to make this right.

This has involved, from my observation, not arguing the toss when really the toss should have been argued. Some of the criticisms are trivial. But it’s understandable that the commission has adopted this strategy.

Killesteyn was terribly unlucky to be the schmuck left standing when the music stopped, but his departure early this year makes the self-flagellation easier. Unfortunately the latest Audit Office update has found fault in the time since Killesteyn.

How much of the bagging of the AEC is warranted? I don’t know enough to answer that, but certainly not all of it. And from what I’ve observed, some JSCEM members and National Audit employees don’t fully appreciate the uniqueness of election management bodies.

As someone once said, elections are the greatest logistic exercise outside wartime.

Australia is actually unusual in having a permanent, comprehensive electoral body like the AEC. In most countries local governments run the vote-taking. This can present problems of partisan interference. Mistakes occur, but usually the blame is cauterised. (Think of Florida’s butterfly ballots in 2000.)

And then there are our vast distances with sparse population.

At an Australian federal election, tens of thousands of employees work in thousands of polling stations across the country. There will be errors. But there can’t be. But they’re inevitable because humans are involved.

So procedures are put in place. And procedures onto procedures. But stuff still happens.

I’ve only read the Auditor-General’s most recent instalment. It no doubt makes some fine points, but also seems to reveal a misunderstanding of the task of an election management body.

(This month’s report contains a couple of mathematical/typographical errors of the sort I’ve never seen the AEC allow through to a published document.

One is: “At the 2013 election more than 91 per cent of the 14.7 million votes counted were received at either a PPVC (18.1 per cent) or static polling place (72.9 per cent).”

No, 13.7 million votes were counted; 14.7 million was the number enrolled.

And more trivially: “fewer than 73 per cent of total votes counted being received at a static polling place.” I make the figure 73.4, which is not “fewer than” 73 per cent.)

The Audit Office has a bee in its bonnet about polling places and the increasing number of people voting before election day.

Last week Brian Boyd (Executive Director, Performance Audit Services Group) formulated an analogy between the AEC and Telstra/Telecom which was revealing of the ANAO’s mindset. He noted that the proliferation of mobile phones over the last two decades had seen telephone boxes all but disappear from our streets, but the AEC by contrast is not keeping up with the modern world and is “still putting a Telecom Australia box on every street corner.”

He was referring to the 7,697 booths around the country on election day last year, with 73, 434 staff.

The comparison is terrible because Telstra is (now) a profit-making organisation. The vast majority of Australians don’t miss phone booths, but a tiny proportion, people without mobiles (including some who don’t have landlines either) sure do. They are few in number and mostly low in income and don’t matter in a business sense.

The AEC’s role is to provide (or try to) voting facilities for everyone, not just most people. If it were a profit-making organisation that was paid by the vote, it would be rational to service only the 70 per cent or so who are easiest to get to. The marginal cost per vote of going that extra mile is very high. Of course.

The auditors make a big deal about the increasing rates of early voting. Modern lifestyles and expectations have led to increasing demands, and a few years ago federal parliament decided to go with the flow and relax the eligibility rules for pre-poll voting.

It had a big effect, and it turns out the AEC overestimated the number of people who turned up on polling day 2013. The ANAO reckons, and from their tut-tutting JSCEM members apparently agree, this led them to man too many polling places and hire too many workers.

But to me it is an example of an electoral body erring on the side of caution. It’s certainly preferable to the opposite: anticipating a big drop in election-day turnout, hiring employees accordingly and running the risk of being understaffed.

If the ANAO and JSCEM nagging succeed in forcing the AEC to anticipate another big drop in 2016, and it doesn’t happen, who’ll be to blame?

The “Telecom” prism leads them astray in other ways.

The latest ANAO report, and JSCEM witnesses, bang on a lot about the decreasing proportion of voters attending on Saturday. They say its decline—to only 73 per cent last year—should mean fewer polling places are needed.

But the proportion has nothing to do with the number of booths and workers required. The total roll grows every year and the facilities needed to cater for, say, 10 million people across the country is independent of what percentage that 10 million happens to be of some larger number.

Now (from ANAO figures) the actual numbers of people who attended polling places did decrease from 10.1 million in 2010 to 9.4 million in 2013. If the AEC’s crystal ball had indeed been operational, what staffing level would have been appropriate?

Even the ANAO accountants would recognise the economies of scale and that the number of booths and staff required wouldn’t drop proportionally. But perhaps they don’t get how unproportional it is.

It might horrify them to know that in 1903, at the first election run by the National Electoral Office, fewer than a million voters were serviced by 26, 254 staff at 4, 525 polling places across the country.

So the staff-to-voter ratio jumped from 1:36 to 1:138 in 110 years, which isn’t bad considering the mechanics, of people lining up, having their names crossed off, given ballot papers etc, has barely changed.

(Whether those ways of doing things should be altered is another topic.)

What would be an appropriate number of booths and staff to take the votes of 9.4 million on polling day? We could look for the last time a similar number turned up. In 2004, the number was 10.2 million. From that year’s annual report it involved 7, 729 polling places and “approximately 65, 000 polling officials”.

(I got 2004 from the 2004-5 annual report. I think I’m comparing apples with apples here, but am not totally sure. There has been a change in what counts as an “ordinary vote”.)

So there were more polling places, but fewer staff, a decade ago.

Maybe the pen-pushers have a small point there.