The Malcolm Mackerras Six And The Question Of How To Define A Landslide

In a weekend newspaper article, the well-known psephologist and election analyst, Malcolm Mackerras, argued that there have been only six federal election “landslide” victories.

In his article, Mackerras nominated the six elections as: 1917, 1929, 1931, 1943, 1966 and 1975.

Mackerras quite rightly objected to the idea that “every second federal election” is a landslide. He described the 2013 election as a “respectable loss” for the ALP but not worthy of being called a landslide.

He said: “However, I have a more rigorous definition, the details of which I have not the space to elaborate now.”

How To Define “Landslide”

I would suggest two essential election statistics as criteria for defining a landslide:

1. The proportion of House of Representatives seats held by the winning party or parties.
2. The national two-party-preferred vote achieved by the winning party or parties.

The primary vote achieved by the election winners is also of some interest but since our system of compulsory preferential voting always provides us with a national figure of combined primary and preferred votes the primary vote alone doesn’t necessarily mean much.

In his article, Mackerras made a crucial point: “…our House of Representatives electoral system is landslide-prone in terms of seats, if not of votes. Such is almost always the case in a system of single-member electoral divisions.”

This is absolutely correct. The Australian system produces exaggerated majorities because of the use of compulsory preferential voting in single-member electorates. Winning parties, by definition, win a greater number of seats by small margins than do losing parties. For example, Tony Abbott’s coalition won 60% of the seats last year but received only 53.49% of the two-party vote and 45.55% of the primary vote.

The Mackerras Six

In classifying his six landslide elections, Mackerras says three of them – 1917, 1943 and 1966 – saw the governments of Hughes, Curtin and Holt re-elected. The other three – 1929, 1931 and 1975 – saw the governments of Bruce, Scullin and Whitlam defeated.

Obviously, the governments in the first group were re-elected and in the second they were defeated, although strictly speaking Whitlam had already been removed by the Governor-General.

Equally obviously, however, an election equation cannot artificially divorce winners from losers. Did Hughes win in 1917 or was the ALP so divided, especially since its leader – Hughes – had defected to the other side, that its election was untenable? Did Holt win in 1966 or was Calwell defeated? Was Whitlam defeated in 1975 or was Fraser chosen to govern?

The 1993 election is the classic example of this puzzle. Did the electorate vote against Hewson or for Keating? Depending on your political standpoint, it’s possible to mount a range of arguments.

It’s a bit like the choice we’re still mulling over today: was the Labor government defeated last year or was the Coalition chosen to lead? Does the act of throwing out one government imply an endorsement of the policies of the opposition? Were we voting in favour of the adults in the room or against the wayward children?

Before we get lost in the philosophy of elections, let’s apply some figures to the Mackerras Six.

The Mackerras Six – The Landslide Elections
Election	Winner	Party/ies	Seats Won In House	Two-Party-Preferred Vote %	Winner’s Primary Vote %
1917	Hughes	Nationalist	53/75 = 70.66%	N/A	54.22
1929	Scullin	ALP	46/75 = 61.33%	–	48.84
1931	Lyons	United Australia/Country	34+16 = 50/75 = 66.66%	–	36.10 + 12.25 = 48.35
1943	Curtin	ALP	49/74 = 66.21%	58.20	49.94
1966	Holt	Liberal/Country	61+21 = 82/124 = 66.12%	56.90	40.14 + 9.84 = 49.98
1975	Fraser	Liberal/National Country	68+23 = 91/127 = 71.65%	55.70	41.80 + 11.25 = 53.05

We can begin to see why Mackerras has nominated these six elections for landslide status. In each of them, the winner took at least 60% of the seats in the House. In two of them, the winner took over 70%.

I’m happy to accept 60% of the House seats as a benchmark for a landslide. On that basis, elections such as 1946, 1987 and 2001 don’t make the grade.

Whilst two-party figures for 1929 and 1931 are not available, and even 1943 involves some estimation, it is clear that the Mackerras Six all had winning two-party votes in excess of 55%.

In each election, the winning party or parties have a high primary vote, ranging from 48.35% to 54.22%. However, the primary vote is largely meaningless because it takes no account of how many votes were garnered by minor parties and independents. The two-party figure is the only meaningful way to compare results from one election to another.

Why Not Include These Elections In The List Of Landslides?

In addition to The Mackerras Six, there are five other elections where the winning party or coalition got 60% or more of the seats in the House: 1925, 1949, 1977, 1983 and 1996.

Should It Be The Mackerras Eleven?
Election	Winner	Party/ies	Seats Won In House	Two-Party-Preferred Vote %	Winner’s Primary Vote %
1925	Bruce	Nationalist/Country	37+13= 50/75 = 66.66%	–	42.46 + 10.74 = 53.20
1949	Menzies	Liberal/Country	55+19 = 74/121 = 61.15%	51.00	39.39 + 10.87 = 50.26
1977	Fraser	Liberal/National Country	67+19 = 86/124 = 69.35%	54.60	38.09 + 10.01 = 48.10
1983	Hawke	ALP	75/125 = 60.00%	53.23	49.48
1996	Howard	Liberal/National	75+18+1= 94/148 = 63.51%	53.63	38.69+8.21+0.35 = 47.25

Some points of note about these five elections:

• The 1925 election was the first official Coalition campaign and resulted in one of the biggest wins in terms of seats and a large primary vote to boot.
• The 1949 election ushered in 23 consecutive years of Coalition government and the nation’s longest-serving prime minister.
• The 1977 election was a virtual repeat of 1975, confirming the electorate’s judgment on Gough Whitlam. From the ALP’s perspective, it was a wasted election and rebuilding the party had to wait until 1980.
• The 1983 election produced one of the most consequential governments in our history and the nation’s third longest-serving prime minister.
• The 1996 election produced the nation’s second-longest serving government and second longest-serving prime minister.

On what basis, I wonder, does Mackerras exclude these elections from his list of landslides?

We might exclude 1949 on the basis of a slender two-party lead. We might exclude 1983 and 1996 on similar grounds in that they don’t meet the 55% two-party-preferred threshold referred to earlier.

Fair enough, I suppose, but all three elections were substantial wins. Menzies became an institution whilst Hawke and Howard had enough fat that they were able to lose ground at their second elections but also go on to win two more elections each.

Surely 1925 and 1977 Count As Landslides?

It’s hard to see why 1925 and 1977 shouldn’t count as landslides. They were both substantial wins for the Bruce-Page and Fraser governments. In both cases, the winners only had one more election in them. Their victories here no doubt helped cushion them against anti-government swings in 1928 and 1980.

Admittedly, 1977 falls just below the 55% two-party mark but it is the third-largest seat count since Federation. Taken with 1975, Malcolm Fraser achieved the largest and third-largest election wins in Australian political history.

If figures were available, the 1925 election probably wouldn’t hit the 55% two-party mark either but the seat count is equal fourth in the combined lists. Having blackmailed the Nationalists into dumping Billy Hughes in 1922, the Country Party quickly made the coalition arrangement an essential feature of the non-Labor side of politics, a coalition that survives to this day.

Numbers Don’t Tell You Everything

Perhaps, as Mackerras may be suggesting, landslides aren’t just about the numbers. As someone who doesn’t believe everything can be measured and who doesn’t believe that statistics explain everything in politics, I’m not altogether opposed to taking into account other factors in defining a landslide, although data should trump dogmatism.

For example, during the 1950s, 1960s and 1970s, one-vote-one value didn’t apply in Australia. The Country Party managed to ensure a 20% tolerance in the allowable variation in enrolments in House of Representatives seats. This ensured they won proportionately more seats than their primary vote indicated. It took Whitlam to reduce the tolerance to 10% and Hawke to legislate a redistribution to implement it.

The Australian Constitution, in Section 24, guarantees each original state at least 5 members. Tasmania would lose one seat if this wasn’t the case. The Constitution guarantees that Tasmania will always be more equal than the other states and that elections are sometimes imprecise measurements of voter sentiment.

Perhaps we should consider whether the election win was achieved by an Opposition coming into government or by a government getting re-elected. Winning office from opposition is undoubtedly the hardest – it’s only happened seven times in the past 60 years – but it’s also more likely to produce a landslide result.

It can be shown that in most elections the incumbent government wins but the landslide re-elections of 1925 and 1977 are rare achievements. Holt’s victory in 1966 was achieved from government but it was also his first electoral contest as leader.

It’s hard to disagree with the Mackerras Six. I’m just inclined to make it Eight and throw in some honourable mentions.

• February 25, 2013: ALP Federal Election Results Since 1910