Four senators were elected with less than a quota of votes. Exhausted votes ranged from a high of more than 9 per cent in NSW to a low of 2 per cent in South Australia.

Voters struggled with large ballot papers.

There was a significant disenfranchisement of many voters in deciding the critical last senator in all states. This may have been due to an emphasis (quite reasonably) by the Australian Electoral Commission on maximising the formal vote by stressing the minimum requirements (six votes above the line, or 12 under it) rather than engaging in complicated reasons why a voter might want to increase the chance of their vote remaining in the count to the bitter end.

The exhausted vote was greater than the difference between the last two candidates in every state. In every state, the result would have been different by one, two or even three senators, if voters had not marked so few boxes that their vote exhausted before the last seat or two were decided.

For example, in Tasmania the last seat was won by the Greens over One Nation by 141 votes, while 9531 valid votes lay idle on the counting-room floor expressing no preference between the two.

What if you could have found those 9531 voters and asked them: "Would you have liked to express a preference between Greens and One Nation for the 12th Senate seat?" My guess is that nearly all would have said yes.

It was worse in NSW, where 414,656 votes (more than a full quota's worth) lay idle on the counting room floor while the Liberal Democrat David Leyonhjelm, One Nation's Brian Burston, the National Party's John Williams and the Christian Democrats' Nella Hall battled it out for the last three seats. Further, the last candidates in four out of six states did not achieve a quota.

The new voting system was supposed to prevent senators being elected on very small percentages of the primary vote. But nationwide we had two senators elected with their party getting less than 4 per cent of the primary vote. And two more whose parties only just got over 4 per cent.

So did Hercules Turnbull clean out the unrepresentative swill in the Augean stables? No. We now have 20 crossbench senators; there were 18 last time.

As it happens, I think this is a good thing. The two major parties together do not deserve more than 56 seats in the 76-seat Senate, by portion of vote or by behaviour. Indeed, they should each have fewer seats in the House of Representatives.

Did the new voting system put the vote in the hands of the people rather than the parties? To some extent yes. To some extent no.

We need to go back to the drawing boards or at least have some tweaking on this hastily put-together Greens-Coalition scheme. But in doing so we might have to accept that several objectives of the voting system are in such conflict with each other that we might never get a flawless system. Rather we might have to balance the aims to achieve something that is acceptable.

Some of the aims are: ease of voting; ease of counting; that those elected are a fair reflection of the voters' will; a fair system of permitting people to stand for election; a system that does not permit undue influence of party factions.

Lessons can be drawn from this election.

There is a direct correlation between the length of the ballot paper and the percentage of the exhausted vote. It is not possible to get ballot papers for each state so let's measure the length of the ballot paper by the number of A4 pages in the AEC candidate counting sheets. For each state with the percentage of exhausted vote in brackets they are as follows: NSW 9 pages (9.23 per cent); Vic 8 (8.58); QLD 8 (7.67); WA 6 (6.28); SA 5 (2.03); Tas 4 (2.81).

So to reduce the exhausted vote we either need to reduce the number of candidates and therefore the size of the ballot paper and/or increase the number of parties or candidates that people vote for – ticking more than six or 12 boxes.

To reduce the number of candidates we could increase the deposit candidates lose if they or their party does not get more than a certain percentage of the first-preference vote.

Or we could have stricter requirements before a group can register as a party, maybe requiring a higher paid-up membership.

There is a conflict here between the ease of standing for election and the ease of voting.

To increase the number of parties or candidates voted for, we need greater education about the system. In the 2016 election, there was simply no time for this. The AEC did a terrific job explaining the new system and getting such a low informal vote. But now we have had one go at it, the finer points can be explained: "To make sure your vote stays in the count to the end you need to number all the boxes either above the line or below the line. You can mark fewer, but your vote might exhaust before the last seat is decided."

There is a conflict here between ease of voting and ensuring voters get a full say. This advice might also encourage people to vote above the line (giving more power to party factions) because people will not be bothered numbering all the boxes below it.

To ease counting and stop candidates with a small percentage of the vote getting elected we could exclude all candidates whose party (or themselves if independent) do not get, say, 4 per cent of the vote. That would have excluded just two of the 72 state senators elected this time with another two just getting over the line. But it would have made the counting so much easier.

Things might be different in a half-Senate election, of course.

In any event, the Senate has returned a result far more reflective of the parties' national vote than the House of Representatives, where the major parties get an easy ride that they are not going to give up easily.

crispinhull.com.au