Monday Melbourne: CLVIII, November 2007
Russell Degnan

Melbourne Central, from the State Library grounds. Taken January 2007

Melbourne Town 27th November, 2007 22:45:22   [#] [0 comments] 

Senate Predictions: South Australia
Russell Degnan

And so finally to the most bizarre race of all. Despite a tendency to vote in high numbers for the Democrats (and in the last election, Family First), South Australia has tended to have a relatively high major party vote, sitting somewhere around 80-85%. The entry of Nick Xenophon, whose last effort at the South Australian state election almost won him three seats in their Legislative Council, means this is very likely to change.

In the past 4 elections, the ALP proportion has been stable at around 42-44 percent, having slowly moved from around 52-53 percent in the early 1980s. With a projected swing upwards of 10 percent however, the polls put the likely result much closer to the latter figures than the former. Morgan predicts (as usual) an outrageous 58% ALP proportion, and Newpoll (in the HoR) a more reasonable 53%. A ten percent swign would put it somewhere near the latter, though much depends on whether it is Labor or Liberal leaking away to Mr X. Notice not only the size of the uncertainty, but also that I predict it to be in the bottom lower corner, with a minor party vote of 29%

A swing of this size means two things. Firstly, that the proportion of the minor party vote needed to win is only 50%. Although Xenophon isn't high on anyone's preference list, he will have sufficient carry to pick up votes from most parties in the middle part of the count, be that from the Greens or Family First or from the major parties. The only danger to him is if (with around 7-9% of the vote) the Greens (with Democrats help or perhaps vice versa) and Family First (with the Liberals or vice versa) sneak above him, subsequently electing both parties on his preferences.

With Xenophon in the count, he will almost certainly take the 5th seat on Family First preferences (if he doesn't pass quote by himself), leaving the Greens, Labor and Liberal. If the Greens come third, Labor will make quota, otherwise the Greens take the final seat. If Xenophon does make quota then the most likely result is a Green victory on Labor preferences, though they may suffer substantial losses to their own vote, allowing Labor or Family First to take the final seat. The most likely result remains a Xenophon, Greens double, however.

Prediction

53% ALP Proportion of Major Party Vote
71% Major Party Primary Vote
2 ALP, 2 Lib, 1 Xen, 1 Grn

Sterner Matters 24th November, 2007 18:22:31   [#] [0 comments] 

Senate Predictions: Queensland
Russell Degnan

While not as bad as New South Wales, Queensland's high minor party vote, and unusual combination of pariah parties with big votes puts those last two spots up in the air. in the past the coalition has been guaranteed three seats in all but 1998 when One Nation and the Democrats shared them. It is likely the ALP will still struggle to capture enough of the vote to compete for a third seat, leaving a really large gap in the middle for other parties to support.

The sheer size of the predicted swing makes accuracy in either that or the minor party vote difficult. Newspoll seems to undercook the Greens vote but have the proportion at 53.8%. Morgan, who normally seem to be overstating the ALP Senate vote, have it at 55.1%. And GetUp, who are seem to overstate the Greens vote, put the ALP proportion at 48.7%. Taking the middle ground, I estimate the proportion at 51.5%, with a relatively modest minor party swing (taking into account the slow decline in the One Nation/Hanson vote) of 4% making it (a still high) 27% overall.

The most significant aspect of the vote being in this region is that neither the coalition nor Labor have strong votes, and are therefore vulnerable to coming third or fourth at the expense of minor parties.

For the coalition (actually, the Nationals), much rests on the size of the Family First and Pauline Hanson vote. If FF can stay in front of Pauline Hanson then they capture her preferences to propel themselves in front of the Liberals for the final seat, locking out Labor whose preferences then elect the Greens candidate. If not, the coalition takes the fifth seat, while Pauline Hanson preferences will either elect the Democrats (with Greens help), or Labor.

An exceptionally large Pauline Hanson vote can also elect Labor and the Greens, or even both the Democrats and the Greens, though in neither scenario are the Nationals more than a few thousand votes from taking the final seat. Liek Victoria and NSW this one is too close to call, but my feeling is that the Nationals and Greens will sneak through.

Prediction

51.5% ALP Proportion of Major Party Vote
73% Major Party Primary Vote
2 ALP, 2 Lib, 1 Nat, 1 Grn

Sterner Matters 24th November, 2007 15:39:25   [#] [0 comments] 

Senate Predictions: Victoria
Russell Degnan

Also relatively stable (with the exception of 1993 when the minor party vote plummeted, Victoria also has a history of resisting big swings, although the polls tend towards something like that today.

Uncertainty around the Greens vote is the major problem making a prediction. Their relative strength is what throws the major party vote around, and a suge to them could dampen the ALP vote. Estimates of the proportion and swing put it somewhere in the range of NSW, though that Green vote means the Senate proportion needs to be shifted towards the Liberal party from the HoR proportion. The predicted ALP proportion here is 53%, and a major party vote of 75%.

As in NSW, this leaves two seats up for grabs, but unlike there, no minor parties seem capable of collecting the necessary preferences. If the Liberals poll very badly, there is a chance for Family First to push for the final seat, but, at the same time, if the Liberals vote falls too far, Family First will still fall short of a full quota. There is an outside chance that the Climate Change Coalition can use first place on the ballot, and then Democrats and Family First preferences to sneak in front of the Greens and then to victory. Sadly too, the Democrats will need a miracle in the form of a 6% vote to overhaul the Greens and use their preferences.

The final two seats, therefore, are a very, very close battle between the Greens (Democrats and other preferences), the ALP, and the Liberals (Family First, DLP, CDP). Most reasonable estimates have one side or the other winning by between a few hundred and a few thousand votes. Predicting this race is practically impossible, however, late movement to the Liberals and/or further dampening of the ALP vote from an increase in the Greens makes me suspect the ALP will just miss out.


Prediction

53% ALP Proportion of Major Party Vote
75% Major Party Primary Vote
2 ALP, 3 Lib, 1 Grn

Sterner Matters 24th November, 2007 14:23:42   [#] [0 comments] 

Senate Predictions: NSW
Russell Degnan

New South Wales is the most stable of all the states, strangely resistant to big swings (1996 excepted) staying within the 45-55% range on primary votes and 15-25% for minor parties. Recent elections have (not surprisingly) been to the rightward end of that spectrum, but only twice has it failed to return a minor party candidate, at the expense of the weaker of the two major parties.

The ALP HoR proportion has traditionally been very similar to the Senate, being within a percent either way. We can therefore get a pretty good idea from recent polling where the likely vote will fall. Or we could, if it weren't so volatile. Digging into the Newspoll makes things a little clearer, which have sat at around 46-41 on ALP primary, much of the extra volatility being the Greens vote. This translates to a 53% ALP proportion, modified 0.5% for HoR difference, a swing of 7%. Adding a swing of 4.5% to the minor parties and we get the following:

Much depends on the right-wing minor party vote in this race. Labor is guaranteed to come in the last 2 or 3, the Greens in the last 3 o4 4, and the Liberals in the last 4 or 5, but that leaves two spots open at the business end.

A number of crucial points late in the count will decide which of three candidates will get up for the fifth seat. The LDP is the main threat, beneficiaries of some ridiculous preference flows and capable of winning with as little as 1/6 percent of the vote, they need to sneak past a series of hurdles to make it. Firstly, the Fishing Party, then Family First. Losses to them can potentially propel the Climate Change Coalition above the Greens and into the fifth seat.

Otherwise, the LDP will suck up CCC votes, putting them up against the CDP, who pick up One Nation, DLP (left of the ALP on the ballot, and therefore potentially upwards of 2%). If the LDP is lower than the CDP, their votes are redistributed (mostly to the Greens), and the higher of the CDP and Liberals will take the fifth seat. A slightly more likely scenario sees the CDP redistribute through the LDP, pushing them in front of the Greens and gifting the final seat to themselves and Labor.

Even that isn't certain however. If the Liberals are excluded before the Greens, their vote will elect both the LDP (or the CDP) and the Greens. A complex count then, probably not decided until a few weeks after the election, not least because BTL votes will matter at several crucial points. Minor parties in the Senate make things interesting though, so I hope for an upset.

Prediction

52.5% ALP Proportion of Major Party Vote
76% Major Party Primary Vote
2 ALP, 1 Lib, 1 Nat, 1 LDP, 1 Grn

Sterner Matters 24th November, 2007 11:13:29   [#] [2 comments] 

Senate Predictions: NT
Russell Degnan

The Northern Territory is at once both the least interesting Senate race, and the most intriguing trend. It is the least interesting because both parties are safely ensconsed over the 33% make needed to make quota. Their NT Senate seat will be the least of their problems if the coalition suffered the type of swing that made it vulnerable.

But it is still interesting, because in defiance of all other states, the Northern Territory slowly shifted towards Labor during the 1980s, and has held that position despite a growing minor party (Greens and Democrats). The latter trend has drawn them close to the national average for the minor party vote. However, the limited number of candidates means that it will be unlikely to go much lower. The graph of past trends is as follows:

Trying to predict the swing up north is practically impossible. Polling is almost non-existent or never reported because the sample is too small, and mostly focused on the marginal seat of Solomon (which isn't showing a strong trend). I'm therefore predicting an 8% broadly in line with polling in South Australia and Queensland. The CLP can rest easy though. Even with the most outrageous swing will the CLP will sneak home on Democrats preferences.


Prediction

56% ALP Proportion of Major Party Vote
83% Major Party Primary Vote
1 ALP, 1 CLP

Sterner Matters 23rd November, 2007 08:26:40   [#] [0 comments] 

Senate Predictions: ACT
Russell Degnan

Just as Western Australia is the most rightward leaning of States, the ACt is the most westward leaning, combining a strong ALP vote (normally 50-60%) with a tendency to vote for the third party (~16%) - previously the Democrats, but now the Greens. And yet,m because the territories only have two senators a piece, the seats have always been split between the two parties. This may no longer be true next week.

The graph of ACT results has generally followed a similar shape to NSW (you'll have to take my word for it), albeit several points across to the left, and with far less diversity on their minor parties. For this reason, the best estimate of the swing is likely to be the same as in NSW. Taking the past three Newspolls as a guide, the swing in ALP major party proportion in NSW is 6.5%. Minor party representation is already high, but has been higher and there are more parties on the ballot than in 2004. A minor party swing of 4.8% again seems reasonable, leavign the graph looking like this:

Two things are worth noting here. The first is that the Liberals are obviously short of a full quota. With only the LDP for preferencing support, the Greens are clear front-runners to win the second seat. The second, is that the ACT has very strong below the line preferencing, so the Liberals may benefit from Labor and Democrats leakage,

The magnitude of the Greens vote makes any other possibility extremely unlikely - though it is possible for Labor win both seats, given an outrageously large swing. This is therefore a straight shootout between the Liberals and the Greens supported by everyone else. The Greens would appear to be favourites, though the graph probably overstates their advantage, given the BTL voting in operation.


Prediction

58.5% ALP Proportion of Major Party Vote
74.25% Major Party Primary Vote
1 ALP, 1 Green

Sterner Matters 22nd November, 2007 23:36:11   [#] [0 comments] 

Senate Predictions: Western Australia
Russell Degnan

Western Australia is the most Liberal leaning of the states, with an astonishing 61% Liberal proportion, and having been somewhere over 53% since 1990. The decline of the Democrat vote has been matched somewhat by the Greens, and those two parties have shared the final seat over the past 6 elections. In 2004, the minor party proportion was substantially lower thnan the recent past, indicating the potential for a reverse next week.

Like Tasmania, the Liberal Senate primary overshot its HoR equivalent by 2.1%, though this is consistent with the past decade, indicating strong independents in the lower house. Polling in WA has been generally favourable to the Liberals, with a predicted swing of between 2 and 5 percent in the HoR. Morgan polling predicted a better result for the ALP in the Senate, at 39 ALP, 42 Liberal, or a 48.1% ALP proportion (major party 81%). However, this would seem to be at the outer limits of the ALP result.

Taking the HoR as a starting point, a 4% swing gives 39% ALP to 44% Liberal. Take 2.5% off the proportion, consistent with 2001 and 2004 (and a high Greens vote) leaves a predicted ALP proportion of 44.5%. A commensurate rise in minor party support, as discussed yesterday, would put the major party vote at 78.5%. The graph of this looks as follows:

The Liberals are not as safe as they might first appear for a third seat. One scenario point to their preferences being cut off and them losing out to the Christian Democratic Party who scoop up a long list of right-leaning preferences (including, crucially, the Nationals) once Eric Wynne (Group M) drops out.

Very few scenarios point to anything other than a Greens victory for the final seat however. It is possible for Labor to come second if the vote is in the far right edge of possible outcomes, but even then, Democrats and the Liberal surplus should be sufficient to get the Greens over the line.

Conversely, when the ALP comes anything but second, their vote (worth from a quarter to half a quota) propels the Greens to the final spot.

Ultimately, the final two seats will go to one right-leaning party (most likely the Liberals, but also the CDP or Family First), and to one left-leaning party (ALP or Greens). There doesn't appear to be any way for the Democrats or other micro parties to garner enough preferences to contest either.


Prediction

44.5% ALP Proportion of Major Party Vote
78.5% Major Party Primary Vote
3 Liberal, 2 ALP, 1 Green

Sterner Matters 22nd November, 2007 02:08:48   [#] [0 comments] 

Senate Predictions: Tasmania
Russell Degnan

Starting then, with the most unpredictable of votes, and the most predictable of Senate races. Tasmania has spent most of the past 24 years defying the national trend, swinging Labor when elsewhere swung Liberal, and swinging to minor parties when elsewhere went for the majors. Until 1996 that is, when it fell nicely into line, voting in a roughly similar fashion to Victoria and NSW.

In the past two elections however, something very odd has been occurring with the major party Senate primary votes. Whereas Labor has 51.5% of the House of Reps. primary vote, they lose 11% of their primary in the Senate. Most of that is to the Greens (who poll almost a full quota), however not all. The Liberals actually recorded 4.1% more in the Senate than the House of Reps. in 2004 (an unprecedented difference). It is hard to see why this would be replicated again, but (along with the lack of any polling) is the major reason why predicting the major party primaries is so difficult. The graph below shows past movements, seat distributions and the prediction for next week.

Morgan polling predicted a massive 59.9%, but an 18% swing in major party primaries would take the result into totally new territory. A more reasonable (if large) 8% swing in HR primaries would give Labor 52.6% to 34%. Swiping off 11% (the same as in 2001 and 2004) for the Green vote makes this 41.6-34, or a 54% ALP primary proportion and 75.6% major party vote.

The Green vote in the last two elections was just short of a full quota, though polls put it as high as 20%. Given Bob Brown and Andrew Wilkie's personal votes, splitting the difference at 17% seems reasonable. Even should they fall short, preferences from WWW (1st on the ticket) and assorted independents will see them over the line. Above 19% the Greens are an outside chance of a second seat on Liberal preferences, but have Family First, LDP and DLP working against them.

The graph shows the likely range for the vote. Because the Greens have a full quota (or near enough to) already, there is only one seat remaining, and the dotted lines become crucial. There are few scenarios where Labor is not one of the last two candidates remaining, as they have close to 2/3 quota on the final seat.

Barring a huge Family First vote, in which case they may overtake them on DLP preferences, the Liberals are unlikely to pick up the last seat without 39% or more.

In the micro stakes, there is an outside chance that: What Women Want could use first place on the ticket (3%) and a Green (4%) surplus to get ahead of the Liberals/Family First and then Labor on Liberal preferences; or that Family First get a sufficiently high vote (5%) to take DLP (2%) and Liberal preferences (7%) and make quota.

Prediction

54% ALP Proportion of Major Party Vote
75.5% Major Party Primary Vote
3 ALP, 2 Liberal, 1 Green

Sterner Matters 21st November, 2007 22:16:24   [#] [2 comments] 

Ratings - December 2007
Russell Degnan


An early start to December, with the India-Pakistan tour following the heels of the Australia and South African summer warm-ups.

Sri Lanka v Australia

Opening Ratings: Aus: 1441.45 Sri: 1100.60
1st Test: Australia by an innings and 140 runs
2nd Test: Australia by 96 runs
Closing Ratings: Aus: 1447.38 Sri: 1095.58

Sri Lankan cricket has been consistent in the past 5 or 6 years. They've been dominant at home - despite a loss to Australia - and weak away - despite wins over England. Their stars are undoubted quality, but until recently the non-Murali section of their bowling line-up hasn't had the fire-power to do much damage. After the World Cup there were hints that this might be changing. Australian pitches are a good leveller in that sense. Fernando and Malinga enjoyed the pace too much; were too short and punished. Maharoof and Vaas were pedestrian, and Murali seems to inspire AUstralian batsmen to prove something. Ignore the small margin in Hobart. It was only that small because Ponting has confidence in his bowlers - not always well placed. The Australian batting abolutely mauled Sri Lanka: 1303 runs for the loss of 11 wickets. Jacques was lucky at times but cashed in. Clarke and Hussey are in superior form, and Symonds and Gilchrist are not players you want to see at 4 for 400. As well as Sangakarra batted in Hobart, and he looked class in both innings, a win there would have been a travesty.

Uncompetitive in the field, the Sri Lankan batting was always going to struggle. They have been over-dependent on Jayawardene and Sangakarra for some time, so it was no surprise to see them bowled out cheaply without the latter, despite Atapattu's efforts. However, they still asked some difficult questions of the Australian bowling lineup. Johnson showed something, despite being impatient. Clark was probably unlucky, not least because he deserves the new ball. MacGill looked second rate, and with an Indian batting lineup shortly to arrive, his test career must be in doubt. The revelation though, 60 tests into a frustrating and unfulfilled career, was Brett Lee: 16 wickets at 17.56. Important wickets too, at key times, combined with tight bowling, especially in Brisbane. A typically expensive and ineffective Lee probably would have meant both tests petered to a draw on flat unforgiving pitches. In the long run, Australia is weaker for losing McGrath and Warne. In the short term, with a firing Lee, they can still boss teams around.


South Africa v New Zealand

Opening Ratings: Saf: 1122.74 NZ: 1065.59
1st Test: South Africa by 358 runs
2nd Test: South Africa by an innings and 59 runs
Closing Ratings: Saf: 1139.37 NZ: 1034.17

Stephen Fleming was always shrewd, and it may be his shrewdest ploy yet to shake off the captaincy when his side was about to go into a serious decline. There are some solid all-rounders in the New Zealand side - Oram, Styris, Vettori - who, as support for an uninjured Bond could, and do, bowl sides out. But the rest of the batting is as inexperienced and untalented as some of the sides Zimbabwe put out before they lost their test status. This was not a pretty series. After conceding over 100 on first innings despite bowling South Africa out for 226, the loss of Bond was the spur for Kallis and Anla to put on 330. Even 46 not out by the new captain couldn't get them to 200 in the second innings. The second test was the same story, except for the bit about bowling out South Africa cheaply, this time New Zealaned conceded 220 to Amla and Kallis, and it was Fleming scoring 54 (their only half century of the series, though Cumming was well set whne he got injured) in a spineless second innings of 136. Steyn was the destoryer, taking 20 wickets at 9.2, but this series bodes ill for New Zealand cricket.


Forthcoming Series:

India (1158.23) v Pakistan (1080.98)- 3 Tests

Yet another series between these two, but disappointingly not the five tests that befit such a contest. India should, after their efforts in England start favourites, but the absence of Sreesanth and RP Singh puts enormous pressure on Zaheer Khan and the spinners. Pakistan are, as ever, enigmas. Much rests on the twin batting talents of Mohammed Yousuf and Younis Khan, and on whether Shoaib Akhtar decides to play (especially without Mohammed Asif). India, at home, would seem to have the edge. Though questions may be asked of their ageing batting lineup and uneven attack, it is hard to see Pakistan makign enough runs to win games.

Sri Lanka (1095.58) v England (1136.24) - 3 Tests

The ratings imply this should be close, as does the English victory in the oneday series that finished before Sri Lanka's whirlwind tour of Australia. However, neither England's batting nor their bowling lacks the bite of a few years ago. This may not matter, if they can contain Sangakarra and Jayawardene, and if their top-order fins some form. But if not, or if the other Sri Lankan batsmen decide to give some much needed support, the improving Sri Lanka will likely roll a weakening England.


West Indies (8th) 833.77
Zimbabwe (9th) 672.64
Bangladesh (10th) 597.99

Idle Summers 21st November, 2007 02:36:55   [#] [0 comments] 

  [next -->]