Friday, June 12, 2015

Would Proposed Senate Reforms Increase The Risk Of A Blocked Senate?

Advance Summary:

1. A previous article on this site showed that proposed Senate reforms to eliminate preference-harvesting disadvantage only preference-harvesters.

2. There apparently remain concerns that the proposed new system would lead to an increased chance of a blocked Senate with the Coalition (or the Coalition plus clearly right-wing crossbenchers) holding half the seats when Labor came to power.

3. Preventing the election of micro-parties off very small shares of the vote does increase the chance for either "side" to from time to time win exactly half of the seats.

4. However, if such a situation does happen, it would be very unlikely to persist beyond a new Labor government's first term.

5. Furthermore, it is only likely to arise in the first place in a case in which Labor is thrashed at one election then wins narrowly at the next (a situation that cannot apply to Labor if it wins the next election narrowly, because of the crossbenchers elected in 2013).

6. Based on the actual votes cast at elections, Labor would actually have had an easier road to passing legislation during its 2007-10 term under the proposed new system than under the current system.


In my most recent coverage of Senate reform (Do Proposed Senate Reforms Advantage The Coalition?) I discussed reported objections to the current JSCEM reform proposal that had been attributed to some Labor Senators.  (Click on the "senate reform" tab for many more articles discussing this, but the proposal is to abolish automatic group-ticket preference transfers between parties, while making it much easier to vote for individual candidates below the line.)  I showed that, if anything, Labor is the party most likely to lose seats it should have won under the present Group Ticket Voting system, and that experience since the rise of the Greens was consistent with that.

At the bottom of the piece I updated with some verbatim quotes from Senator Sam Dastyari, and some rather feral speculations about what on earth could be motivating the seemingly inexplicable opposition to reform from some within the ALP.  I've since had it suggested that there may be one specific concern motivating Labor hesitance about adopting the proposed reform, and it's one that's quite consistent with Sen Dastyari's comments about about "forever preventing a progressive Senate".  
The reported concern is that the proposed reform would make it too easy for the Coalition to obtain a blocking majority (exactly 50% of the seats), such that the ALP in government would not be able to do anything.  This might happen if every state returned a 3 Coalition, 2 Labor, 1 Green result.

Before I look in fine detail at the risk of this happening, I should point out that a somewhat increased chance of blocked Senates (with 50% of the seats held by the Coalition, and 50% by Labor and the Greens combined) is an unavoidable "risk" - under the current system - of getting rid of Senators who get in on small percentages of the vote.  In elections in the last few decades, there have fairly often been Senators outside the major or established minor parties elected more or less on their own right rather than preference deals or other problems of the current broken system (Harradine, Xenophon, One Nation in 1998 and PUP in 2013 in Queensland).  However, some elections have seen all seats go to the Coalition, Labor and The Greens and /or Democrats.   

What I will question here isn't that this could happen, but the assumption that Labor would be likely to have to face it in government.  But first a little history ...

The Ghosts of 1974-5

The Australian Senate has had three very distinct phases.  Until 1949 the use of a "slate" system frequently resulted in the same party winning every seat in a given state.  From 1949 the introduction of proportional representation ended this, but the requirement to number every box without any savings provisions for errors led to huge informal voting rates (in many elections around 10%) that appeared to disadvantage Labor.

The Dismissal crisis of 1975 was significantly caused by a blocked Senate (once one Independent had joined the Coalition), which became an artificial Coalition majority after the Queensland government refused to accept Labor's nomination for a casual Senate vacancy.  The Whitlam government had won the 1974 double dissolution election narrowly in the Lower House (51.7% 2PP), but in many states failed to convert Senate primary vote leads to results better than a 5-5 split of right vs left/centre.  The ALP's performance relative to the Coalition was about a point weaker in the Senate than in the Reps, possibly because of high informal voting in the Senate. Labor lost one seat in Queensland by less than that margin, but for which it would have held 30 seats and been able to pass any legislation supported by moderate crossbencher Steele Hall.  The effectively deadlocked result of the Senate election was thus most likely a combination of a pretty close election anyway, the system disadvantaging Labor, and bad luck.

The next major Senate voting reforms came into force at the 1984 election.  The introduction of Group Ticket Voting greatly reduced unintended informal voting, but did so at the cost of creating the opportunity for micro-parties to harvest preferences.  As noted before, until 2013, micro-parties were only very rarely successful in this, though from time to time a major or minor party would win a seat it didn't really deserve.

Another important reform that came in for 1984 was the increase in the size of the Senate.  A major impact of that was that at half-Senate elections, each state now elected six Senators instead of five.  This change made 3-3 left-right splits in at least some states more or less inevitable, and in theory there could be six of them and hence a blocked Senate.  While this change has been criticised for that reason, it can also be praised on the grounds that it is no longer possible for a clearly less popular side of politics nationally to control the Senate by winning 3-2 splits in the four least populous states.

But how often has the Coalition done well enough to be in a position to secure a blocking majority?

Seats won since 1990: current vs proposed

The 1984 election saw seven Senators elected per state and the 1987 election was a double dissolution, so these are not very useful for assessing what might happen in half-Senate elections with six Senators elected per state.  To look at them historically we have just the nine Senate elections from 1990 to 2013 to go on.

These were the Senate seat results of those elections.  In each case I also give the Lower House outcome, and the Lower House 2PP as a crude indicator of the relative strengths of the major parties:

40 seats are decided at each Senate election.  I've highlighted cases where either the Coalition or Labor plus the Greens (but not plus the Democrats, who more commonly worked with either side) won at least half the seats at a given election.  Two consecutive such results gives that side at least a blocking majority with at least 38 of 76 seats (the four Territory seats are re-elected every election.)  The Coalition won at least half the seats three times (1996, 2001, 2004) but only obtained a Senate majority in the 2004-7 term.  Since the Greens became a major force, Labor and the Greens have won a combined majority in both the 2007 and 2010 elections, resulting in a Senate majority for the two combined in the 2010 term.

Based on these results it is hardly worth even talking about a blocking majority after the 2016 election, whatever the system, if it is a half-Senate election.  The 2013 hangover would mean that the Coalition would need 23 seats, or Labor plus the Greens 24, requiring either side to win four-seat blocks in at least three states.  (It is possible Labor and the Greens will someday win a 2-0 block in the ACT, which would leave them needing three 4-2s rather than four to get up to 24).   The issue is what happens once the proposed new system has been in for at least two half-Senate elections in a row.

The following are estimates of the seat breakdown had all Senate votes cast between 1990 and 2013 been cast under the proposed new system.  (I assume that WA 2013 would not have been invalidated, and that the ballot paper size issues leading to the huge Liberal Democrat vote in NSW 2013 would not have happened.)

In forming these estimates, I've assumed the winners would have been the same except where there is a fairly clear case based on the primary votes that they would have been different.  The differences for 2004-2013 were discussed in the previous article (I've not altered the lineball ones).  I've also reassessed Queensland 2004 - I think the Coalition would have probably got four up under the proposed new system in that case as well.

From 1990 to 2001 the differences between the two systems based on the votes cast were probably few.  While the Democrats and the Greens were both polling significant votes, there were cases where one would overhaul Labor on the preferences of another, and I've given one Dems seat in 1990 and one Greens and one Dems seat in 1993 to Labor.  I also doubt Pauline Hanson would have been overhauled based on the votes cast in Queensland in 2001.

In this reconstruction the Coalition would have won a blocking majority result only in elections that it won, so there is no support there for the scenario of Labor coming to power and facing a blocked Senate.  According to this simulation, based on the votes actually cast, Labor in the 2007-10 term would have been able to pass legislation supported by Nick Xenophon and the Greens.  In reality they needed the support of Xenophon, the Greens, and hostile Family First Senator Steve Fielding.  It is strange that having come to government with an almost blocked Senate under the current system, some in Labor are concerned about the proposed new system possibly blocking the Senate.

It can also be seen that the Coalition only won 20 or more seats in this reconstruction when it won, and even then not always.  It would be possible for it to win 40 seats at two consecutive elections including a victory followed by a defeat, but neither of the ways this could realistically happen (21-19 or 20-20) are all that probable - and both could happen under the existing system if micro-parties stopped winning as many seats as in 2013.

Those adept at looking for the catch will notice that my simulation almost gives Hanson the balance of power in the 2004-7 parliament.  However, that is based on the assumption that the system itself would not have altered voter behaviour.  It's all but certain that it would have.  The  minor-party left would have adopted tactics to counter the rise of One Nation under the proposed system, such as a Greens-Democrat deal for one party to run hard in one state and the other in another.  It's possible even that had we had the proposed new system in place, the nasty transition from Democrats to Greens as the main third party in the Senate would have played out very differently.

In terms of the chances of the Coalition getting a blocking majority at any given election, it's possible to compare its Lower House performances with actual Senate seats won, and with those that might have been won under the proposed system, based on actual votes cast:

The relationship between Lower House performance and seats won in the Senate is much stronger under the new system based on the votes cast, because the influence of micro-party preference aberrations is weaker, making 2013 no longer an outlier.  While the Coalition would have won (and did win) an outright majority in the 2004-7 parliament based on its 2001 and 2004 results combined, on average it would need a 2PP of 53% across two elections to win a blocking majority.  If one of those results is a loss, then that becomes rather unlikely.

As I mentioned in the previous edition, it's likely that the right generally would come up with some strategies to counteract its apparent slight disadvantage under the proposed new system.  This could include running more separate Liberal and National Senate tickets (a double-edged decision since if they got it wrong, exhaust would hurt them), and also right micro-parties merging to form parties capable of getting, say, 7% in some states and fighting for a seat.  The latter strategy would more often take seats from the Coalition than anyone else, but every now and then the right would get an extra seat out of it.  Even so, if it did happen that Labor came to office facing a blocked Senate, that would most likely result from the party being heavily beaten at one election and then narrowly winning the next.  (The opposite has been more common in practice, though there is no obvious reason for that.)

It is worth explaining why a blocked Senate would rarely result for an incoming Labor government.  At any time there is generally at least one crossbench Senator somewhere - typically in a smaller state - who is not clearly either left or right, and who either party could work with on a reasonably regular basis.  So for the losing side to block the Senate it usually needs to win more Senate seats than the side that won the most recent election.  But because of the variation in Labor support between states, good results for Labor will usually result in a 4-2 left-right split (3 Labor, 2 Liberal, 1 Green) somewhere (especially in Tasmania), while 4-2 right-left splits like in Queensland 2004 are rare.  It isn't even necessary for the left to get a 4-2 split to avoid a blocked Senate so long as there is a 3-2-1 split in its favour involving a crossbencher somewhere.

Also, there is no evidence that an alternative reform model (such as a primary vote threshhold) would avoid all potential for a blocked Senate.

And If There Is A Blocked Senate?

My suggestion is that neither party should be afraid of the prospect that the Senate may sometimes be blocked when they take office from Opposition.  This has been a reasonably common situation in Australian federal political history, with almost half of all new federal governments coming to power with a more or less blocked Senate.  It is frustrating, but a first-term government usually has considerable electoral goodwill and can deal with a blocked Senate either by forcing a double dissolution or by gaining seats at the next half-Senate election.

If a new system frequently led to a government facing a blocked Senate even after winning two elections in a row, that would be a greater cause for concern.  Because of the very malapportioned nature of the Senate (in which small states have as many seats as large ones), and the granular nature of Senate results, it's impossible to say for certain that this couldn't ever happen, but it's highly unlikely.  If it did ever happen, then it would be cause for a far greater reassessment of how we elect the Senate than just the question of translating votes into seats within each state.

Some level of risk of a blocked Senate for a new government will arise under any system that combines the result of two elections.  Short of ceasing to use a two-election cycle there is no way around this.  It would be a more serious problem if blocked Senates persist past the first term of a new government, but there is no evidence that this would ever be so.

It's true that the new system will sometimes produce a blocked Senate in some circumstance in which the current system would instead have produced an unmanageable mess of semi-randomly elected crossbenchers, but there is no reason to accept claims that the new system would risk "forever preventing" a progressive Senate.  Nor is the risk of Labor coming to office with a blocked Senate in the future as great as some may think.


  1. Would expanding the Senate to 14 seats per state help?

    In this way, 7 Senators are elected per state at each election, making it more likely that one side or the other gets a working majority. Instead of endless 3/6 dead heats, the winning side would have a realistic chance of winning 4/7, alone or with like-minded minor parties. The problems of "permanent blocking majorities" by the losing side would be reduced.

    The lower quota with 7 seats as opposed to 6 would help mollify the minor parties who are upset with the current preference changes. It's harder to win a seat if you're just preference harvesting, but easier to win if you're a genuine minor party who pulls a decent primary vote.

  2. Having 14 senators per state would require adding about another 26 MHRs as well to maintain the nexus between the houses, with a huge redistribution like in 1984. But yes, in theory an odd number of Senators per state reduces the chance of a blocked Senate, since to win exactly half the seats a party will probably need to win more than half (eg 4/7) in exactly 50% of the state contests over two elections.

    The risk, as I mentioned with the old 10-seats-per-state system, is that we someday get a big-states-vs-small-states situation where a party wins the Reps by virtue of its performance in Vic and NSW but the smaller states give its opponent a lasting outright Senate majority.

    1. 18 MHRs... territory senators do not count in calculating the size of the House (see Mackellar v Commonwealth). - JD

    2. Would 18 senators per state still suffer this problem, or would the quota for a half senate election (10%?) be low enough such that there would always be enough minors to stop either side getting 5/9 seats?

      The 18 senators didn't come out of nowhere (kind of), it's more a property of "The Australia Institute" suggesting an increase in the lower house to ~223 MPs, back when there was a ruckus about NT about to revert to one lower house seat

  3. what about 18 senators a state? Would that still have a big-states-vs-small states situation, or would the half senate quota be low enough to push a bunch of minors in?

    18 senators came to mind because "The Australia Institute" suggested increasing the lower house in order to stop the NT from losing their first seat in a submission to APH

    1. 18 Senators per state would make either a Coalition or a Labor-Greens blocking majority quite difficult - on current vote shares neither "side" wins any 5/9 slates. The interesting thing with applying 18 seats to the 2019 results however is how well One Nation do. A party that consistently gets about 4% per state does almost as well in half-Senates with an 18 seat system as one that consistently gets 10%. I get the 2019 state numbers as L-NP 23 ALP 17 Green 6 PHON 6 UAP 1 JLN 1, and 2016 as L-NP 22 ALP 18 Green 6 PHON 4 NXT 2 DHJP 1 JLN 1, so there would actually be a combined Coalition-PHON majority.


The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.