There are few things sexier in this universe than a discussion of Senate electoral mathematics, and psephologists across the country couldn't help but be a little excited in the downstairs department over the discussion of the lie of the land in the Senate after FedElec04.
The piece by former Greens staffer Ben Oquist in Crikey last week is a good starting point, and Oquist argues that the coalition only need to repeat their 2001 Senate results this time around, and they will have 38 of the 76 Senate seats. If the Libs get over the line in the House of Reps, they will fall short of a Senate majority after they appoint one of their number as the Senate President (37 of 75). Should the ALP win in the lower house, and therefore need to provide the President, then the Lib/Nats would have a Senate majority (38 of 75), and hence an effective veto on all bills. Such a situation is unusual in a proportional representation system, but occurred in Victoria during the first term of the Bracks government, when a single-member electorate system was in place.
The general rule when it comes to determining who will win the 6 Senate seats for each state at each election is to presume that they will split with 3 to the conservative parties (Libs, Nats, One Nation, Harradine) and 3 to the progressive parties (ALP, Democrats, Greens, Murphy, Lees). The territories will split one each along similar lines. Any deviation from this is most unusual.
In 2004, this general rule is likely to hold true. Fortunately for the coalition, with Harradine and One Nation unlikely to make an impact, they have the conservative 20 seats all to themselves. The ALP, however, will need to contend with the Greens, who, like the Democrats before them, will feel quietly confident of picking up a seat in almost every state. If this is to hold true, then in the next parliament the coalition would indeed have 38 Senate seats, and the power to block supply if in opposition. The Libs would never do that. In the Senate. Surely not.