THE MACKERRAS PENDULUM – AND ME – MALCOLM MACKERRAS

For decades, Malcolm MacKerras’ electoral pendulum has been part of predicting the outcomes of Australia’s elections. His pendulum for the federal election of 2025 was published along with his 600 word piece in The Daily Telegraph on Saturday 25 March 2025. What to make of it in these days of weekly polling from multiple sources? Malcolm Mackerras addressed The Sydney Institute on Wednesday 23 April 2025 to discuss his contribution to Australia’s psephological studies and the development and popularity over time of his “Pendulum”.

THE MACKERRAS PENDULUM – AND ME

MALCOLM MACKERRAS

It came as a pleasant surprise that The Sydney Institute would invite me to give a talk with this title at this time. While I have always described the Mackerras Pendulum as “the original and the best”, it never really occurred to me that anyone would take my bragging seriously at my age. If I had been asked to give a talk in which I set my own title it would have been headed “My agenda for Australian electoral reform”. I explain below why that would have been the title I would have chosen.

The best objective description of the Mackerras Pendulum came in 2016 when Professor Murray Goot of Macquarie University published in the Australian Journal of Politics and History (Volume 62, pp 59-86) a scholarly article titled “The Transformation of Australian Electoral Analysis: The Two-Party Preferred Vote – Origins, Impacts and Critics”. I quote from part of his first two paragraphs on page 59:

Since the early 1970s, when Malcolm Mackerras popularised the concept, “the two-party preferred vote” (2PP) has eclipsed the first preference vote as the way electoral support in Australia is measured. The 2PP, whereby second (and sometimes subsequent) preferences for the minor parties, micro parties, and Independents are counted as either votes for Labor or votes for the main non-Labor parties is the currency in which most discussions about party support are now transacted, notwithstanding that these votes were cast against major parties. Contrary to the claim that there is no “distinctive Australian approach to the study of elections and electoral behaviour”, there is: this is it.

Although the idea of a 2PP long pre-dates Mackerras, he popularised it. Where single-member, lower house electorates are likely to be won by Labor, Liberal or National Party candidates, as overwhelmingly they are, the 2PP has become the standard way of translating votes into seats – most famously via the Mackerras pendulum. . .

Then, on page 60, Goot notes that “Mackerras first promoted the 2PP in a 1971 paper devoted to the meaning of electoral swing” and then refers to the paper I gave at a seminar in March 1971 to the Department of Political Science at the Australian National University. The title of the paper was “Electoral ‘Swing’: What does it mean?”

Although my quest to be famous appears to have begun in March 1971 my interest in matters psephological began 20 years earlier when, in March 1951 (at the age of eleven) I noticed that the Governor-General, Mr William McKell, had granted a double dissolution to the then prime minister, Mr Robert Menzies. That excited my interest with the effect that I closely followed the April 1951 general election. I can, therefore, claim to have followed in detail the elections of 1951, 1954, 1955, 1958, 1961, 1963, 1966 and 1969 which I call my “pre-pendulum” elections.

The 21 federal general elections from 1972 to 2025, therefore, constitute my “post-pendulum” elections. It is a fact of my history that I have been able to persuade newspapers to publish my pendulum for all of the 21 most recent federal elections so that my very recent bursting into print in News Limited papers for Friday 28 March 2025 records me as claiming: “This Mackerras Pendulum is my 21^st for federal elections published by newspapers. The first was in December 1972.”

Back in 1951 virtually every federal electoral division recorded a straight contest between a Liberal (or Country Party) candidate and a Labor candidate, and the margin between the two was referred to as the winner’s “majority”. Thus, in the most perennially swinging seat of Bass (Tasmania) the votes were recorded as 16,469 for the sitting Liberal Bruce Kekwick and 14,346 for Labor’s Colman O’Byrne. So, Kekwick’s majority was 2,123 votes. In the southern Sydney seat of Barton, the votes were 20,949 votes for the sitting Labor member Herbert Vere Evatt and 20,706 for the Liberal candidate Nancy Wake, so Evatt’s majority was 243 votes. At that election there were unopposed returns in Darling Downs (Queensland), Hindmarsh (SA) and Kalgoorlie (WA).

At the following election in May 1954 essentially the same happened. There were unopposed returns in Angas (SA), Bradfield (NSW), Hindmarsh (SA), Mallee (Victoria), Murray (Victoria), Richmond (NSW) and Wentworth (NSW). Back in those days I recognised Joan Rydon as being Australia’s leading psephologist and she is also referred to by Goot in his description that the 2PP “long pre-dates Mackerras”. She found a way to account for unopposed returns but never thought of the term “two-party preferred vote”. In her early works Rydon always referred to “the two-party vote” but she later changed that to her “adjusted two-party vote”.

Australia’s leading psephologists

It is my considered opinion that, from September 1940 to March 1971, Joan Rydon was Australia’s leading psephologist. From March 1971 to March 2010, I was Australia’s leading psephologist and since March 2010 that title has passed to Antony Green. The choice of the date of the first transition is clear from the comments on my first page. The choice of the second transition date is that Antony turned 50 in March 2010. He has been employed as a psephologist by the ABC since August 1989.

There is a difference between Joan Rydon and the two men named above. She never received any civilian honour for her work. For reasons described in more detail below my suggestion is that there should be a street in a new outer Melbourne suburb to acknowledge her name. As one who believes in diversity, equity and inclusion, I approve of giving female names where appropriate.

There are two Australian men who have been made an Officer in the Order of Australia (AO) for his contribution to Australia’s democracy. Mine was given on Australia Day 2006 and the citation is: “For service to the community by raising public awareness of and encouraging debate about the political process in Australia and other Western democracies, and through commitment to reform and improvement of the electoral system, and to education.” Antony was given his AO in 2017, and his citation is: “For distinguished service to the broadcast media as an analyst and commentator for state and federal elections, and to the community as a key interpreter of Australian democracy.”

The transition from “majority” to percentages

On the first day of July 1956, two senators from the then Democratic Labor Party took their seats in the Senate chamber. The days of a very strict two-party system were over. Minor parties entered the Senate consequent upon the electoral reforms implemented by the Chifley Labor government in 1948, effective from the December 1949 general elections. For the 1949, 1951, 1954, 1955 and 1958 general elections, however, it was still sensible to speak in pre-Mackerras terms. However, the 1961 election changed everything. It was no longer sensible to speak of the DLP as a party opposing the Liberal Party. It became effectively a partner in the conservative coalition against Labor.

John Howard describes that election best. In his book The Menzies Era he has a chapter 13 titled “Saved by Santamaria”. The chapter begins by describing the fact that in 1961 the Menzies government appeared to have been saved by the victory of Jim Killen in the Queensland electoral division of Moreton. Howard then describes the mythology of that. On pages 282 and 283 he sets the record straight. He writes:

The real preference story of that election was the faithful flow of DLP second preference votes to Liberal candidates in Victoria. Communist preferences in Moreton didn’t save Menzies; DLP preferences in Victoria did. If they had not so strongly favoured Menzies, the count in Moreton would have been academic – except for Jim Killen.

The results in 1961, 1963, 1966 and 1969 convinced me of the need to change Australian forms of analysing elections. In March 1971 that challenge was met by my presentation of the paper cited above.

Having explained the two-party preferred vote, I now explain the pendulum and what made me think of it. In 1953 I began my practice of buying books with election statistics in them. For Australia, I bought the eleventh edition of the Parliamentary Handbook and Record of Elections. It gave me the statistics for 1946, 1949 and 1951, some of which I have quoted above. However, the United Kingdom also had a general election in 1951, and I bought the book The Times House of Commons 1951. It gave me all the statistics I needed in respect of the general election occurring on 25 October 1951. That election was dominated by straight contests between Conservative and Labour candidates. For example, in the West division of Walthamstow Labour’s candidate Clement Attlee secured 23,021 votes while the Conservative candidate Edward Du Cann received 11,447 votes, so the Labour majority is shown on page 104 as 11,574 votes.

By the way, as an aside, that was the election when Attlee lost the office of prime minister to Winston Churchill. Six years earlier, in 1945, Attlee had enjoyed a landslide victory over Churchill. Attlee’s seat then was Stepney which was abolished in a later redistribution. His vote then was 8,398 and the Conservative candidate Lieutenant Woodward secured 1,618 votes, so Attlee’s majority was 6,780 votes. The absurdity of saying that his majority nearly doubled is illustrated by the fact that Attlee secured 83.85 per cent of the vote in 1945 and “only” 66.79 per cent in 1951.

The above information excited my interest in Dr David Butler (1924-2022) the eminent British psephologist who devised for the BBC election coverage a diagram known as the “swingometer”. It showed numbers of seats falling for each one per cent of swing. I quickly realised the difference between the British House of Commons and the Australian House of Representatives. The British House of Commons in 1951 had 625 members while the Australian House of Representatives had only 121. So, whereas there were far too many seats for the British to include in a simple diagram Australia had chosen to have a number small enough to enable every seat to be fitted into a single pendulum.

The cancellation of deviations

I first met David Butler in 1967 when he was in Canberra on a sabbatical leave. Essentially, I owe the Mackerras Pendulum to him. He explained to me that such a diagram could be sold on the basis of his observation that both in the UK and in Australia overall swings were essentially uniform. However, that would not mean perfect uniformity. There never would occur an election in which every seat would record exactly the same swing – but the doctrine would still be useful because the deviations from uniform swing could be relied upon to cancel out.

The first election of the publication of the Mackerras Pendulum saw the doctrine work perfectly. In December 1972 Labor secured a gain of eight seats on a national swing of 2.5 per cent. If the swing had been perfectly uniform Labor would have gained these eight seats: Phillip (Liberal, NSW), Hume (Country Party, NSW), Evans (Lib, NSW), Griffith (Lib, Qld), Lilley (Lib, Qld), Herbert (Lib, Qld), Denison (Lib, Tas) and Mitchell (Lib, NSW).

In the event Labor failed to gain Griffith and Herbert. But it gained the seats named below which needed much bigger swings: Cook (Lib, NSW), Macarthur (Lib, NSW), Holt (Lib, Vic), Casey (Lib, Vic) La Trobe (Lib, Vic) and Diamond Valley (Lib, Vic). But Labor also lost these seats: Sturt (SA), Forrest (WA), Bendigo (Vic) and Stirling (WA). So, Labor gained exactly the number of seats it would have gained on a perfectly uniform swing. The deviations cancelled out.

The federal elections of 1974, 1975, 1977, 1980, 1983 and 1984 went by and I succeeded. Not only did I get my pendulums published in newspapers and The Bulletin magazine, but I was also always able to explain that the doctrine of cancellation of deviations worked. Meanwhile, various other analysts tried alternative pendulums and/or other diagrams to explain the same phenomena. They failed. I was able to explain that the Mackerras Pendulum was the original and the best.

Then came the disaster of 1987. Not only did I make a fool of myself on television I finished the year needing to explain why the doctrine of cancellation of deviations had failed so miserably. It was contrary to every doctrine I had proclaimed to have a government increasing its majority while the overall swing was against that government. The result was that in May 1988 I produced a booklet titled Double Dissolution Election: Statistical Analysis: July 11, 1987. I ask permission now to quote myself in these paragraphs on pages 15, 16 and 17:

In the aftermath of the election there was an enormous amount of controversy explaining the paradox of the election. I mean by that the fact that Labor increased its majority from 16 to 24 seats, notwithstanding an overall swing of one per cent against it. Furthermore, Liberal-National needed a uniform swing to win of 2.3 per cent following the 1984 election but now needs 3.3 per cent following the 1987 election. John Howard took the Opposition forward by one per cent on one way of calculation but backward by one per cent on another way of calculation. The received opinion among journalists is that Labor lost ground in its heartland (safe Labor seats) where it could afford loss of votes but cleverly, skilfully and professionally “targeted” the marginal seats where it performed well. In my opinion that view does not correctly explain what happened. I think a look at the regional pattern of swing is the most useful thing to do. From Table 7 it can be seen that Labor lost ground in “Labor Australia” while gaining ground in “split Australia”. This is a useful division of the country for this purpose. However, it is even more useful to sub-divide “Labor Australia” into two so that, electorally, we think of there being three regions as follows:

Region 1. This is the Newcastle-Sydney-Wollongong conurbation where every seat recorded a swing to Liberal, the average being 3.2 per cent. It consists of all the seats in New South Wales except the 14 genuinely country seats of Calare, Cowper, Eden-Monaro, Farrer, Gilmore, Gwydir, Hume, Hunter, Lyne, New England, Page, Parkes, Richmond and Riverina-Darling. It is worth noting that the Newcastle-Sydney-Wollongong conurbation, at 37 seats, has more voters than the whole of “split Australia” (30 seats).

Region 2. This is the remainder of “Labor Australia” and includes the 14 country seats of New South Wales (listed above) plus the Australian Capital Territory, Victoria, South Australia and Western Australia.

Region 3. This is “split Australia”, already defined. . .

The table appears to confirm the impression that Labor lost most in the safest Labor seats. However, it should be pointed out that safe Labor seats were disproportionately located in the region (Region 1) which recorded the greatest swing against Labor anyway. Of the 38 safe Labor seats, 19 were in Region 1, 17 were in Region 2 and two were located in Region 3.

The paradox of the election is best explained by regional variation, accompanied by a close examination of the pre-election pendulum. The most striking feature of that pendulum is the fact that six of the ten most marginal Liberal-National seats were located in Region 3.

There was a 1.5 per cent pro-Labor swing in Region 3. A perfectly uniform swing of 1.5 per cent in Region 3 would have yielded Labor six gains from Liberal-National. And that is exactly the number of gains Labor made in Region 3. And, of course, Labor gained nothing outside Region 3.

There was a 0.9 per cent anti-Labor swing in Region 2. A perfectly uniform swing of 0.9 per cent in Region 2 would have cost Labor one seat – Chisholm (Vic). And that is exactly what happened. The status of Chisholm changed from being the most marginal Labor seat on the pre-election pendulum to being the most marginal Liberal seat on the new pendulum.

The absence of Labor marginals in Region 1 was the key to the election. The average swing to Liberal was 3.2 per cent. But Labor was lucky in that it had only two seats in that region which would fall in such a swing, Barton and Lowe. It lost Lowe.

After discussing whether Barton was held by “targeting”, I then went on to ask the question: “Why did Labor lose most votes in Region 1?” The answer I give is: “Three reasons, I think. Rents, rents and the Unsworth government.”

It has been my practice following each election to analyse whether the pendulum worked or not. By that I mean: was there a reasonable level of cancellation of deviations? My conclusion federally is that it worked in 1972, 1974, 1975, 1977, 1980, 1983, 1984, 1990, 1993, 1996, 2001, 2004, 2007, 2010, 2013, 2016, 2019 and 2022. It failed in 1987 and 1998.

My assessing of my own work is only useful if I enjoy third party validation. In that regard Murray Goot, emeritus professor of politics at Macquarie University, is my best friend. Consequently, I placed on my website two articles by him published in The Australian newspaper during the 2022 campaign. The first was titled “It’s time to flick the switch to the national vote”, published on Thursday 21 April. The second was titled “Who will win? It’s all in how the pendulum swings.” It was published on Thursday 5 May.

So, who will win in 2025?

To answer the big question of the May 2025 general election, I began by doing what I always do. Based on opinion polling I made a guess at the distribution of the national two-party preferred vote. I then applied that to my pendulum. After that I gave attention to individual seats. It is all set out in an article by me published in The Daily Telegraph for Friday 28 March 2025, the day the House of Representatives was dissolved. The pendulum and the article are published on pages 8 and 9. In Melbourne The Herald Sun, In Brisbane The Courier Mail and in Adelaide The Advertiser carry the same article and pendulum though with minor differences in headings and size of pendulum. The article opens with this:

At Christmas, Anthony Albanese will still be Prime Minister. He might lead a majority Labor government, but I think a minority one more likely. I lack confidence in all predictions, but I am still willing to make these: 71 seats for Labor, 65 Liberal-National plus 14 on the crossbench of which three will be Greens ad 11 independents.

The accompanying Mackerras Pendulum is my 21^st for federal elections published by newspapers. The first was for the December 1972 election. In those days Australia had a simple two-party system making it easy to explain a pendulum based on such a concept…

I now ask myself this simple question: “If you are badly wrong, please explain”. Here is my explanation: I made a wrong guess at the overall distribution of the two-party preferred vote. I would now be extremely surprised if Albanese were not Prime Minister at Christmas time. My prediction now is that the number of seats for Labor will be 76 with 45 Liberals, 16 Nationals plus 13 on the crossbench of which two will be Greens and eleven independents. In other words, the nearly four weeks of campaigning between Friday 28 March and Wednesday 23 April have gone badly enough for the Liberal Party and the Greens as to cause me to take four seats from the Liberal Party and one from the Greens. So, I think Labor will win five more seats than I thought a month ago. That has converted my prediction of a minority Labor government into a majority Labor government.

For those who are really interested in details benefit may be gained from reading my Switzer Daily article posted on the morning of Thursday 17 April, Maundy Thursday. It is titled “Will Anthony Albanese be our PM after May 3?”, the title given to it by the editor, not me. My suggested title was “Anthony Albanese predicted to prevail”. Anyway, that article goes into considerable detail as to how I arrived at those numbers. However, I did not go into speculating about the future of Perter Dutton if my numbers turned out to be correct. I do that now.

My guess is that Dutton would remain Leader of the Opposition. His supporters would say something like this: “Dutton was a good leader who was unlucky in the way events turned out. In any event he made a net gain of four seats for the Liberal Party (from 41 to 45) and increased the Coalition’s share of the two-party preferred vote from 47.9 per cent in May 2022 to 48.6 per cent in May 2025. By May 2028 he will have learnt fully the truth of John Carrick’s aphorism: ‘You cannot fatten the pig on market day’ and will be able to benefit from such learning.”

My future – and that of my pendulum

At the age of 85 my psephological efforts are now concentrated on electoral reform and I have a strange ambition: I hope to be able to name myself to have been Australia’s greatest non-politician electoral reformer since Catherine Helen Spence (1825-1910) who was an advocate for female suffrage and proportional representation. Australia’s greatest electoral reformers have been politicians beginning with Andrew Inglis Clark (1848-1907) but there have been about a dozen non-politicians who have helped to promote electoral reform. It is worth noting that at the three consecutive May federal elections (2019, 2022 and 2025) there have been federal electoral divisions of Spence in South Australia and Clark in Tasmania – and their boundaries on 3 May 2025 will be identical to those of the two previous back-to-back May elections.

There is a belief I have that there exists a class of worthy citizens who are never recognised in their lifetime with any kind of honour, but who deserve recognition. In my opinion such a man was Bogey Musidlak (1953-2017) who died suddenly and unexpectedly in August 2017 at the age of 63. I was invited to give a tribute to him at his funeral. I told the 15 people present at the service that he was a much under-estimated man, and I would approach the ACT Public Place Names Authority to see if a street might be named after him in a new Canberra suburb. After some five years of negotiating, I received a notification that my nomination had been successful. In the new Canberra suburb of Denman Prospect there are now a dozen names given to men and women commemorated under the nomenclature theme Activism and Reform. The description of Musidlak Rise is under the heading “electoral reform” and reads as follows:

Boguslaw (Bogey) Musidlak was a long-term President of the Proportional Representation Society of Australia (1994-2017) and a leading Australian electoral reformer. He campaigned strenuously in 1992 for legislative elections in the Australian Capital Territory (ACT) to use the single transferable voting (STV) system. STV voting, known in Australia as the Hare-Clark system, is a proportional and direct election, candidate-based system. Musidlak’s leadership, advocacy and persistence is regarded as instrumental to the installation of the Hare-Clark system in the ACT. Likewise, Musidlak’s influence was important in achieving an important innovation to the ACT Hare-Clark system; that a single first preference beside the name of a candidate is a valid, formal vote. The Hare-Clark system replaced the initial modified d’Hondt system employed upon the introduction of self-government to the ACT in 1989.

My current position is that I nominate Musidlak to have the title “greatest non-politician Australian electoral reformer since Catherine Helen Spence.” My ambition is to replace him for that title. Just as I replaced Joan Rydon for the title “Australia’s leading psephologist” my ambition now is to replace Musidlak with this other title. In the meantime, there should be a Rydon Street (road, crescent, circuit, rise or place) somewhere in Melbourne. There is another comment I make in the meantime. I never agreed with his innovation to the Hare-Clark system. However, I went along with it so that we could present a united front to get the Hare-Clark system adopted for elections to the Legislative Assembly for the ACT.

As for the future of my pendulum I predict it will go out of existence when I die. At present I am the only psephologist who gets his pendulums published in newspapers. There are other pendulums around, but they are not really pendulums. They are listings of seats which the blogger in question calls his “pendulum”. Meanwhile The Australian has sacked me and has chosen to re-work election statistics into a “tower of power” which I think to be an idiotic diagram. Readers who want to see the current “federal tower of power” can find it in The Weekend Australian for 29-30 March.

There is one ambition I do not have. I don’t want a Canberra street to be called after me. In any event there is already a Mackerras Crescent in the Canberra suburb of Theodore. The nomenclature theme of Theodore is Civilian War Effort during World War I and World War II. Edward Granville Theodore (1884-1950) was once a Queensland premier and later deputy prime minister. For this purpose, however, he is described as “Director-General of Allied Works 1942-44”. Two pages after that description comes this:

MACKERRAS CRESCENT: Ian Murray Mackerras (1898-1980), scientist; served AIF World War I; Director Queensland Institute of Medical Research 1947; with CSIRO first in 1928 and then again in 1961; primarily responsible for successful control of malaria and scrub typhus; published The Insects of Australia, 1970; received many honours for his work.

Uncle Ian received many academic honours in his lifetime, but he otherwise was not properly recognised, in my opinion. In that he was like Bogey Musidlak but unlike me. I have been properly recognised in my lifetime which, I submit, is far better that being recognised when one is deceased.

The irritating Antony Green

My successor with the title “Australia’s leading psephologist” is Antony Green – but he irritates me in two ways. The first is in relation to electoral reform. We seem to find each other on the opposite side of every debate, and all I can do is point out that the citation for my AO refers to my commitment to reform and improvement of the electoral system where his citation makes no such mention. In practice, my ability to point that out is of no use to me – except to my ego. Politicians are not interested in that. They pursue the electoral interests of the party that gave them their seats.

Justifying my pedantry

Therefore, it is in relation to the pendulum that I deal for my concluding remarks. I am the man who invented the pendulum, and I begin by saying that my key statistic has always been the percentage swing needed by the candidate or party who lost last time to win at the next election. For example, take page 49 of my book Australian General Elections, published in 1972. The Cronulla-based seat of Cook (NSW) is shown as yielding this 1969 final count: 26,861 votes (52.81 per cent) for Liberal Party sitting member Don Dobie and 24,001 votes (47.19 per cent) for the Labor challenger Cliff Mallam. Therefore, I show this entry: “Swing required to lose – 2.9 per cent”.

If I were writing that today I would say the Liberal Party holds Cook with a 2.9 per cent buffer. Antony Green, by contrast, would say that the Liberal Party holds Cook on a 2.8 per cent “margin”. My argument then was (and today would continue to be) that if 2.8 per cent were shaved off the Liberal percentage then the Liberal Party would still have the seat with 50.01 per cent of the vote. Unfortunately for me, my fame has greatly diminished, and Antony’s has grown hugely, so he has a much bigger fan club. Consequently, pretty much everyone these days uses the term “margin” to describe what is really the percentage swing required – to which I attach the word “buffer” for short.

My current pendulum has two versions. There is the original – beautifully drawn and printed out by the Brisbane firm Digital Synergy Pty Ltd of 26 Navigator Place Hendra. And then there is the “New Microsoft Word Document” which is the way the pendulum was published in those four News Ltd papers that have taken it. The original has all the details on the reverse side that a newspaper would not have the space to publish. For example, it has all the buffer statistics for cases which were once rare but now quite common. These seats now are Bradfield, Canberra, Clark, Cooper, Cowper, Curtin, Fowler, Goldstein, Grayndler, Groom, Indi, Kennedy, Kooyong, Mackellar, Mayo, Nicholls, Sydney, Wannon and Wills.

The reverse side is headed “Fitting nine parties into a two-party system”. For the most part it is an essay discussing the seats that I have just mentioned. However, it also discusses the two rival words “margin” and “buffer”. It shows the actual final 2022 statistics in three key marginal seats, Gilmore (NSW), Lyons (Tasmania) and Sturt (SA). All three are unchanged in boundaries, Lyons and Sturt because they are divisions in states not needing any redistribution. Gilmore is in a state where the boundaries did change but Gilmore itself was left unchanged.

Take Gilmore first. The figure for Labor winner Fiona Phillips was 56,039 votes or 50.17 per cent. The Liberal loser Andrew Constance secured 55,666 votes or 49.83 per cent. My comment is “Swing needed by Liberal: 0.2 per cent”. In that case my “buffer” and Antony’s “margin” are the same.

Then take Sturt. The Liberal winner, James Stevens, scored 56,813 votes or 50.45 per cent while Labor’s Sonja Baram secured 55,797 votes or 49.55 per cent. My comment is: “Swing needed by Labor: 0.5 per cent”. To find Antony’s statistic I went to his listing of seats on his blog which he calls his pendulum. His “margin” is also shown as 0.5 because he has rounded up 50.45 to 50.5. If he had rounded that down to 50.4 then his “margin” would have been 0.4.

A very important seat in this election is Lyons in Tasmania, a seat whose result may well determine whether Labor gets 76 seats or “only” 75. The Labor winner, Brian Mitchell, secured 37,341 votes or 50.92 per cent while the Liberal loser, Susie Bower, won 35,997 votes or 49.08 per cent. My comment is: “Swing needed by Liberal: 1.0 per cent” By contrast Antony’s “margin” is 0.9 per cent. However, I would say that if the swing were only 0.9 per cent to Liberal then Labor would be left with 50.02 per cent and would keep the seat.

A seat where the 2025 result is not in the slightest doubt is Mayo (SA). On the reverse side of my Digital Synergy pendulum, I show the final vote for incumbent independent Rebekha Sharkie as 72,355 and that for 2022 Liberal challenger Allison Bluck as 43,859. The percentages are 62.26 for Sharkie and 37.74 for Bluck. Green says: “Sharkie’s margin is 12.3 per cent”. I say: “Sharkie’s margin is 24.5 per cent and her buffer is 12.3 per cent.” The words in print read: “Swing needed by Liberal: 12.3 per cent.”

In my essay “Fitting nine parties into a two-party system” I do keep the word “margin” with this comment: “To me the margins are 373 votes in Gilmore, 1,016 votes in Sturt and 1,344 votes in Lyons.” So, I shall continue to be pedantic. For the semi marginal Labor seat of Hawke in Victoria others can follow Antony and say: “Held by Sam Rae on a margin of 7.6 per cent”. See The Australian for Wednesday 2 April, page 5. I say that Sam Rae’s buffer is 7.7 per cent. See the Herald Sun for Friday 28 March, pages 6 and 7. However, I also have a table “Battleground Victoria” which shows the precise statistic of 7.62 per cent, that being the AEC calculation for a seat changed in boundaries. Below that is my article “Minority rule the likely outcome but it’s still looking like Albo will have more to celebrate”. Perhaps more importantly there is an article by staff journalists titled “Victoria key in road to victory”.

I should mention that I engage in three forms of personal pedantry – but the other two do not relate to my pendulum. Therefore, this paper creates the need for me to justify this particular piece of pendulum pedantry. I do that now.

In addition to doing Australian pendulums for all our lower houses that are elected from single-member electoral divisions, I do an American pendulum for their presidential elections. I take as my example the swing state of Michigan at the November 2020 election. The votes cast were 2,804,040 for Joe Biden which was 51.41 per cent of the two-candidate vote. For Trump the votes were 2,649,852 or 48.59 per cent. Therefore, I wrote that the swing needed by Trump in 2024 was 1.5 per cent. Were Antony Green to have done a pendulum he would have said: “Biden’s margin was 1.4 per cent”. The Michiganders, however, along with all American analysts, universally said that Biden’s margin was 154,188 votes or 2.82 per cent. In other words, the American thinking is the same as mine. The margin is the difference between the winner and the runner-up candidate. Therefore, my parting comment is to say: “What’s good enough for Michigan is good enough for me.”

On Tuesday 1 April 2025 there were special elections for the US House of Representatives in two rural Florida congressional districts. These two special elections were what we call by-elections. The first district was vacant because Republican Matt Gaetz had resigned. The sixth was vacant because Republican Mike Waltz had resigned. Both resignations were caused by the expectation of appointment to the Trump cabinet, successful in the case of Waltz, unsuccessful in for Gaetz. To show the relevance of the above to my pendulum, I quote from an article by Bruce Wolpe titled “Is this the beginning of the end for Trump? Count the reasons”. It appeared on page 32 of The Sydney Morning Herald for Saturday 5 April:

The first voting since November shows a weakened Trump position. In Wisconsin . . . In Florida, there was a special election for two open House seats. Trump carried those districts by 30 points in November. The Republicans won, as expected, but their margin was cut in half – a bad sign for next year’s House races.

Wolpe is a Sydney resident, but he is also the longest standing American commentator in Australia. When he writes that Trump carried both districts by 30 points, he means Trump 65 per cent and Harris 35 per cent. Members of Antony Green’s fan club, by contrast, would mean Trump 80 per cent and Harris 20 per cent. For me all I can say is that what is good enough for the Floridians is good enough for me. That is why I insist on using the word “buffer” and reject using the word “margin” when the statistic really refers to the percentage swing needed for a seat to change hands.