31/10/2024
The Effect Optional Preferential Voting had on the 2023 NSW State Election
19 April 2023
Preamble
Preferential voting gives a more democratic system and avoids the electoral distortion of ‘First Past the Post’ which favours those with bigger primaries and can give huge distortions of the vote, such as occurs in the UK where there is an effective north-south division between Labour and the Tories, and a very polarised system, as non-major parties get far fewer seats than their share of the primary vote.
NSW has ‘optional preferential voting’ which is an anomaly in Australia, which Federally and in other States have compulsory (or mandatory) preferential. This simplifies the voting system, but also distorts the results towards a ‘first past the post’ system. It was introduced by Neville Wran in 1980 as part of some excellent reforms, but it might be wondered if the less democratic Optional Preferential was introduced to make it difficult for the Liberals and Nationals when they had ‘3 cornered contests’, i.e. Liberal, National and Labor. If so, it is backfiring now, as on this analysis optional preferential voting has given NSW Labor six fewer seats and an extra 3 Independents would have been elected in the 2023 election.
This paper looks at some seats and tries to quantify what difference a change to the compulsory preferential system would have made to the result of the NSW election.
This analysis suggests that Labor would have gained an extra 6 seats, and Independents an extra three. Labor would have gained Drummoyne, Goulburn, Holdsworthy, Oatley, Ryde and Terrigal. The Independents would have gained North Shore, Pittwater and Willoughby, (though WIilloughby would have been already Independent after the 2022 by-election.)
Method
Where a candidate gets an absolute majority, clearly they have won the seat. If the final margin of victory is less than the exhausted vote, there is a possibility that if the votes had not exhausted, the result may have been different. The question is which candidate would the exhausted votes have favoured. A reasonable guess is that the voters who exhausted would have voted the same way as the voters who filled in all squares for their primary vote.
If this is applied to a number of electorates with close margins, a reasonable estimate of the effect of compulsory voting can be made.
This paper looks at some electorates that have final victory margins less than the exhausted vote number, and multiplies the ratio of known preferences for the last two candidates by the number of exhausted votes. This is then compared to the current margin of victory to see if the result would have been different. These results can then be examined to see the change in the composition of the Parliament that would have resulted from a change to compulsory preferential voting. The seats that would change below are in alphabetical order.
The record for each seat shows the number of Formal Votes (FV), the Final Margin of Victory (FMV) and the number of votes that Exhausted (E).
The tables for each seat show the results for the successful candidate and the second candidate. They show the Primary Vote count (PV), then the Final Vote (FV) count after preferences have been distributed. The difference between these is the number of preferences that the candidate received. The Total Distributed Preferences Distributed (TPD) are added and the ratio between them determined as an index of how the preferences flowed. These ratios are then multiplied by the number of Exhausted Votes to see what the final count is likely to have been if preferences were mandatory.
(A more detailed description of the tables is in an Appendix at the end if it is required).
Drummoyne
Total Formal Votes 51,639. Final Margin of Victory 1,285. Exhausted Votes 2,308.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Di Pasqua Lib | 24,526 | 25,308 | 782 | 0.1691 | 390 | 25,698 |
| Little Labor. | 20,182 | 24,023 | 3,841 | 0.8308 | 1918 | 25,941 |
| FMV = 1,285 | TPD=
4,623 |
E= 2,308 | New Result
-243 |
Comment:
If there were compulsory preferential voting, Labor would have won this seat. It is remarkable how few preferences flowed to the Liberals- only 16.9%. 83% of the Exhausted votes would have gone to Labor, winning it the seat.
Goulburn
Total Formal Votes 50,775. Final Margin of Victory 1,170. Exhausted Votes 5,575.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Tuckerman Lib | 20,737 | 23,185 | 2,448 | 0.3804 | 2,121 | 25,306 |
| Pilbrow Labor. | 18,028 | 22,015 | 3,987 | 0.6196 | 3,454 | 25,469 |
| FMV = 1,170 | TPD=
6,435 |
E = 5,575 | New Result
-163 |
Comment:
Goulburn would have changed hands from the Liberals to Labor if there had been compulsory preferential voting.
Holsworthy:
Total Formal Votes 48,791. Final Margin of Victory was 331 Votes. Exhausted Votes 4,404.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Ayyad Lib | 20,449 | 22,359 | 1,901 | 0.4092 | 1,802 | 24,161 |
| Maroney Lab | 19,284 | 22,028 | 2,744 | 0.5906 | 2,602 | 24,630 |
| FMV = 331 | TPD=
4,645 |
E = 4,404 | New Result
|
Comment:
The result is that with compulsory preferences the 2nd candidate, Maroney of Labor, would have won by 469 votes.
North Shore
Total Formal Votes 48,177. Final Margin of Votes was 3,658, The Exhausted Vote was 6,808.
| Primary V | Final V | Pref Rec’d | Ratio of Pref | Exh. V Added | New Total
FV + EVA |
|
| Wilson Lib | 21,308 | 21,987 | 1,901 | 0.0802 | 547 | 24,161 |
| Conway Indep | 10,308 | 18,329 | 7,802 | 0.9199 | 6,261 | 24,630 |
| FMV = 3,658 | 8,481 | E = 6,808 | New Result
|
Comment:
The result is that the Independent, Conway would have won with compulsory preferential voting. Note that the primary of Conway was 21.9%, less than half of Wilson at 44.2%, but the preference flows to Conway were 7,802 votes, 92% and Wilson 697, only 8%.
Ironically, the current 2 Candidate preferred, which favours Wilson by 55.69% to 44.31%. makes it look like a safe seat, but this is because of the Exhausted Votes, as the existing system looks only at the Final Vote Count.
Oatley
The total formal vote was 50,196. Final Margin of Victory 754. Exhausted Vote 3,032
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Coure Lib | 22,877 | 23,959 | 1,082 | 0.244 | 740 | 24,698 |
| AmbihaipaharLabor | 19,851 | 23,205 | 3,354 | 0.756 | 2,292 | 25,497 |
| FMV = 754 | TPD=
4,436 |
E = 3,032 | New Result
-799 |
Comment:
The result would have been that Labor would have won. The heavy preference flows suggest that a majority of people do not want the candidate that was elected.
Pittwater
Total Formal Votes 49,511. Final Margin of Victory 754. Exhausted Votes 3,387.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Amon Lib | 22,137 | 23,365 | 1,228 | 0.1970 | 667 | 24,032 |
| Scruby Indep | 17,754 | 22,759 | 5,005 | 0.8030 | 2,720 | 25,479 |
| FMV = 754 | TPD=
6,233 |
E = 3,387 | New Result
-1,447 |
Comment
Again, the Independent, Scruby would have won if there were compulsory preferential voting because the Liberal got only 19.7% (667) of the Exhausted Preferences. On the preference ratio Scruby would have got 80.3% (2,720). Since 754 + 667 = 1,421 is less than 2,720, Scruby would have been elected by 1,299.
Ryde
Total Formal Votes 53,840. Final Margin of Victory 54. Exhausted Votes 3,032.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Lane Liberal | 24,383 | 25,431 | 1,048 | 0.1933 | 586 | 26,017 |
| Howison Lab. | 21,004 | 25,377 | 4,373 | 0.8067 | 2,446 | 27,822 |
| FMV = 54 | TPD=
5,421 |
E = 3,032 | New Result
– 1,804 |
Comment
This seat had a re-count and aroused a great deal of interest.
Again, the result would be changed and Howison of Labor elected if there were compulsory preferential voting, because the majority of preferences would favour Labor. Few preferences went to the Liberals who in this case were greatly helped by the lack of preference flow due to Optional preferential voting.
Terrigal
Total Votes 50,470. Final Margin of Victory 1,467. Exhausted Votes 2,703.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| Crouch Liberal | 23,507 | 24,467 | 960 | 0.2107 | 569 | 25,036 |
| Boughton Lab. | 19,703 | 23,300 | 3,597 | 0.7893 | 2,134 | 25,434 |
| FMV = 1,467 | TPD=
4,557 |
E = 2,703 | New Result
– 398 |
Comment: Boughton of Labor would have won if there were compulsory preferential.
Willoughby
Total Formal Votes 52,830. Final margin of Victory 2,450. Exhausted Votes 5,826.
| Primary V | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh V Added | New Total
FV + EVA |
|
| James Lib | 23,032 | 24,727 | 1,695 | 0.1711 | 997 | 25,724 |
| Penn Indep. | 14,064 | 22,277 | 8,213 | 0.8289 | 4,829 | 27,106 |
| FMV = 2,450 | TPD=
9,908 |
E = 5,826 | New Result
-1,382 |
Comment:
Penn, a well organised Independent had massively more preferences than James, and would have beaten him with compulsory preferential voting. In contests with a strong 3rd candidate, in this case Labor, the preferences are crucial.
Conclusion
If current political trends to minor parties continue, a change in the voting system to compulsory preferential is likely to make an even greater difference in future and this is likely to favour both minor parties and Labor.
Appendix:
Definitions:
1st (Successful) Candidate’s Primary Vote = 1CP
2nd Candidate’s Primary = 2CP
After distribution of preferences
1st Candidates Final Vote = 1FV
2nd Candidates Final Vote = 2 FV
Final Vote Margin of Victory FMV = 1FV – 2FV
Exhausted Vote = E
If E > (1FV – 2FV), then the seat must be considered as able to have had its result changed by having the Exhausted vote preferences considered if preferential voting were compulsory.
The number of preferences P which went to each of the last 2 candidates are:
Ist Candidate 1FV – 1CP = 1P
2nd Candidate 2FV – 2CP = 2P
Thus Total Preferences Distributed 1P + 2P = TPD
The ratio of preferences received will be:
1st Candidate = 1P/TPD
2nd Candidate = 2P/TPD
So if the Exhausted votes followed the primaries of those who gave preferences the extra votes received by the candidates (EVA) would be:
1st Candidate 1EVA = 1P/TPD x E
2nd Candidate 2EVA = 2P/TPD x E
If the Exhausted Votes Added (EVA) to the second candidate is greater than the Final Margin of Victory, FMV (=1FV – 2FV), then compulsory preferential voting is likely to have changed the results. This is noted as a New Result, with a negative sign if there is a change to the existing result.
So the key to the tables becomes:
| Primary Vote | Final Votes
FV |
Pref Rec’d | Ratio of Pref | Exh. V Added = Ratio x E | New Total
FV + EVA |
|
| Successful 1st Candidate | 1CP | 1FV | 1P | 1P/TPD | 1EVA | 1FV + 1EVA |
| 2nd candidate | 2CP | 2FV | 2P | 2P/TPD | 2EVA | 2FV + 2EVA |
| FMV = Final Margin of Victory = 1FV – 2FV | TPD= Total Preferences Distributed = 1P + 2P | Result =
(1FV + 1EVA) – (2FV + 2EVA) |