Before the election, the distribution of seats in the Australian Senate was:

• 34 Liberal/National Coalition (right to center-right)
  
• 31 Australian Labor Party (ALP) (center-right to center-left)
  
• 9 Australian Greens (center-left to left)
  
• 1 Nick Xenonphon, South Australian Independence (center-right, mebbe?)
  
• 1 Democratic Labor Party (center-right)
  

So if the Greens voted with Labor, that was 40 of 76, and if the Greens voted with the Coalition that was 43 of 76, so the Greens held the balance of power.

After the election, the makeup is now:

• 33 (-1) Liberal/National Coalition
  
• 25 (-6) Australian Labor Party
  
• 10 (+1) Australian Greens
  
• 1 Democratic Labor Party
  
• 1 Nick Xenophone (SA)
  
• 1 (+1) Family First Party (SA)
  
• And four new parties:
  
  • 2 Palmer United Party (right to center-right)
    
  • Liberal Democratic Party (NSW)
    
  • Australian Motoring Enthusiasts Party (VIC)
    
  • Australian Sports Parties (WA)
    
  
  

So the Greens have lost the balance of power ... but its not as if it was really "their fault". If they had repeated their success of the previous election, they would have gained more seats, but they did gain a seat compare to the 2007 Senate result that they were defending.

By contrast, while the ALP lost six (after losing one in the previous election). It was the loss of the Labor seats that cost the Greens the Balance of Power, since now if the Greens vote together with the ALP, that is a total of 35, and 37 is required for a majority. By contrast, if the Coalition picks up four out of the eight minor party and independent votes, they can pass legislation through the Senate without Green votes at all, though perhaps at the cost of some baubles for key constituencies in a few states.

In the Australian proportional preferential system for election to the Senate, voter can vote "below the line" for a large number of individual candidates, or "above the line" for a single party. In NSW, for example, there were 110 Senate Candidates, so the ballot paper is quite large and even so polling places are equipped with magnifying glasses. One has to cast votes for at least 90% of the candidates with no more than two breaks in sequence for it to be a formal, vote, so there is an incentive to vote above the line.

When voting above the line, preferences flow as dictated by the party's preferences as lodged with the Australian electoral commission, and minor parties engage in substantial negotiation over how they will preference the votes cast for them. These preference deals can not only affect the outcome among the major and 3rd parties, but if a minor party is lucky, they can stay ahead of the cut, receiving preference flows from eliminated parties as they go, and make it into the ranks of the 3rd parties. This is exactly how the Australian Motor Enthusiast Party and the Australian Sports Party made it into the Senate.

I have, indeed, seen some suggestions that the growing use of tactical preference deals cost the Greens Senate seats in this election, but my analysis, below, is that it was rather the swing in the primary vote toward center-right and right candidates that did the damage.

The way that my analysis proceeds is by comparing the actual election result to the seats won with complete quotas and then for the balance to the parties receiving the largest partial quotas. When there is a departure from that, I look at the final distribution that elected the sixth senator from that state to see what preference flows had switched the order of parties in terms of primary vote.

Analyzing Senate Election Results compared to Primary Vote

Queensland: 3 Libs, 2 Labor, 1 Palmer United Party:

• 2Lib & 2Lab on full quotas, then the Libs 0.8565, PUP 0.7236 ... so the final result was exactly in line with the primary vote

NSW: 3 Libs & Nats, 2 Labor, 1 Dems (Liberal Democratic Party):

• 2 Lib&Nat full quotas, 2 Labor full quotas, partial quotas: Dems 0.6222, Greens 0.5446, Lib&Nats 0.4249, Palmer 0.2470, the Greens lost the last seat on preference distribution, which was primarily from the HEMP Marijuana Legalization Party and the Wikileaks Party both preferencing the Dems for tactical reasons rather than support for party's issue.

Vic: 2Lib, 2Lab, 1Greens, 1 Australian Motoring Enthusiast Party:

• On primary vote, 2 full quotas for Libs and Labor, then partial quotas Greens 0.7791, Libs/Nats 0.7464, Lab 0.3207, Palmer 0.2647, so it was the Libs/Nats (I'm guessing the Nats) done out on preferences ... that was on minor party preferences across the board (from Sex Party through Palmer United) putting AMEP ahead of the Libs/Nats.

Tassie: 2Lib, 2Lab, 1Greens, 1PalmerUnited:

• On primary vote, Libs and Lab 2 full quotas, then Greens 0.8156, Libs 0.6059, Palmer 0.4850, so again it was the Libs done out on preference in favor of Palmer, with in the final distrbution, only Sex Party preferences flowing to the Libs over the Palmer United party.

South Australia: 2Libs, 1Lab, 1Greens, 1Xenephon, 1Family First:

• Primary vote 1 each Libs, Labs, Xenephon. Partial quotas Libs 0.8680, Xenephon 0.8117, Labor 0.5945, Greens 0.4956, Family First 0.2639, Dems 0.2447, Palmer 0.1902. Here it was Labor and Xenephon done out by the Greens and Family First on preferences. (Note that this result sees Family First returning to the Senate, and AFAIR their previous seat was held as a result of advantageous preference flows)

WA: 3Libs, 1 each Labor, Australian Sports Party, Greens:

• Primary vote: 2 full quotas Libs, 1 full quota Labor, then Labor 0.8890, Libs 0.7648, Greens 0.6598, Palmer 0.3667, Nationals 0.3012, Dems 0.2394, Australian Christians 0.1138,  Sex Party 0.1003. Here is was Labor done out by Ozzie Sports Party.

The Impact of Preference Deals

It looks rather like the net damage was to the Lib/National Coalition losing two seats to 3rd and Minor Party conservative candidates and Labor losing one seat to a Minor Party conservative.

The first impact is likely to be relatively minor. Many of the controversial policies that the Coalition may wish to pursue will be consistent to the positions of minor party conservative Senators, so the price to win their vote is unlikely to be high,

The second impact may have been quite a bit more substantial, since 26 ALP Senators and 10 Green Senators would be half of the Australian Senate, which is not enough to pass a measure, but is enough to deny a measure a majority.

As far as the Greens themselves and preference flows, what they lost on the swing they picked up on the roundabout. Since the ALP governed under a negotiated deal with the Greens, it is not surprising that they were unable to repeat their success of 2010, but now that they are again in the position of Independent Opposition, they may well be able to gain back the ground required to retain their 2010 seats in the next Senate election.

But for the most part, it looks to me that it was not preferences so much as a rightward swing of the primary vote that was responsible for the swing in the composition of the Australian Senate.

So, the Greens did not lost seats ... indeed, they picked up one Senate seat and held their Lower House seat. But the Balance of Power position is not a position that a party holds on its own: its held by virtue of the relative strength of the other parties vying for your support. As it turned out, the 31 ALP Senators were the "ground" on which the Green Balance of Power was standing ... and by virtue of the Labor Party's abysmal performance in this Senate Election, the Green Party lost that ground.