:Later-no-harm criterion

{{Short description|Property of electoral systems}}

class="wikitable sortable floatright"

|+ Voting system

!Name

Comply?
Plurality

|Yes{{NoteTag|Plurality voting can be thought of as a ranked voting system that disregards preferences after the first; because all preferences other than the first are unimportant, plurality passes later-no-harm as traditionally defined.|name=p}}

Two-round system

|Yes

Partisan primary

|Yes

Instant-runoff voting

|Yes

Minimax Opposition

|Yes

DSC

|Yes

Anti-pluralityNo{{cn|date=November 2024}}
ApprovalN/A
BordaNo
DodgsonNo
CopelandNo
Kemeny–YoungNo
Ranked PairsNo
SchulzeNo
ScoreNo
Majority judgment

|No

Later-no-harm is a property of some ranked-choice voting systems, first described by Douglas Woodall. In later-no-harm systems, increasing the rating or rank of a candidate ranked below the winner of an election cannot cause a higher-ranked candidate to lose. It is a common property in the plurality-rule family of voting systems.

For example, say a group of voters ranks Alice 2nd and Bob 6th, and Alice wins the election. In the next election, Bob focuses on expanding his appeal with this group of voters, but does not manage to defeat Alice—Bob's rating increases from 6th-place to 3rd. Later-no-harm says that this increased support from Alice's voters should not allow Bob to win.

Later-no-harm may be confused as implying center squeeze, since later-no-harm is a defining characteristic of first-preference plurality (FPP) and instant-runoff voting (IRV), and descending solid coalitions (DSC), systems that have similar mechanics that are based on first preference counting. These systems pass later-no-harm compliance by making sure the results either do not depend on lower preferences at all (plurality) or only depend on them if all higher preferences have been eliminated (IRV and DSC), and thus exhibit a center squeeze effect. {{Cite journal |last=Lewyn |first=Michael |date=2012 |title=Two Cheers for Instant Runoff Voting |url=https://papers.ssrn.com/abstract=2276015 |journal=6 Phoenix L. Rev. |language=en |location=Rochester, NY |volume=117 |pages= |ssrn=2276015 |quote=third place Candidate C is a centrist who is in fact the second choice of Candidate A’s left-wing supporters and Candidate B’s right-wing supporters. ... In such a situation, Candidate C would prevail over both Candidates A ... and B ... in a one-on-one runoff election. Yet, Candidate C would not prevail under IRV because he or she finished third and thus would be the first candidate eliminated |via=}}{{Cite journal |last=Stensholt |first=Eivind |date=2015-10-07 |title=What Happened in Burlington? |url=https://ideas.repec.org//p/hhs/nhhfms/2015_026.html |journal=Discussion Papers |language=en |pages=13 |quote=There is a Condorcet ranking according to distance from the center, but Condorcet winner M, the most central candidate, was squeezed between the two others, got the smallest primary support, and was eliminated.}} However, this does not mean that methods that pass later-no-harm must be vulnerable to center squeezes. The properties are distinct, as Minimax opposition also passes later-no-harm.

Later-no-harm is also often confused with immunity to a kind of strategic voting called strategic truncation or bullet voting.{{cite web |last1=The Non-majority Rule Desk |date=July 29, 2011 |title=Why Approval Voting is Unworkable in Contested Elections - FairVote |url=http://www.fairvote.org/why-approval-voting-is-unworkable-in-contested-elections |accessdate=11 October 2016 |website=FairVote Blog}} Satisfying later-no-harm does not provide immunity to such strategies. Systems like instant runoff that pass later-no-harm but fail monotonicity still incentivize truncation or bullet voting in some situations.{{Cite journal |last1=Graham-Squire |first1=Adam |last2=McCune |first2=David |date=2023-06-12 |title=An Examination of Ranked-Choice Voting in the United States, 2004–2022 |url=https://www.tandfonline.com/doi/full/10.1080/00344893.2023.2221689 |journal=Representation |language=en |pages=1–19 |arxiv=2301.12075 |doi=10.1080/00344893.2023.2221689 |issn=0034-4893}}{{Cite journal | last=Brams |first=Steven |title= The AMS nomination procedure is vulnerable to 'truncation of preferences' |year=1982 |journal=Notices of the American Mathematical Society |volume=29 |pages=136–138 |issn=0002-9920 |publisher=American Mathematical Society |url=https://www.ams.org/cgi-bin/notices/nxgnotices.pl?fm=main¤t=198202}}{{Cite journal |last1=Fishburn |first1=Peter C. |last2=Brams |first2=Steven J. |date=1984-01-01 |title=Manipulability of voting by sincere truncation of preferences |url=https://doi.org/10.1007/BF00119689 |journal=Public Choice |language=en |volume=44 |issue=3 |pages=397–410 |doi=10.1007/BF00119689 |issn=1573-7101}}{{rp|401}}

{{TOC limit}}

Later-no-harm methods

The plurality vote, two-round system, instant-runoff voting, and descending solid coalitions satisfy the later-no-harm criterion. First-preference plurality satisfies later-no-harm trivially, by ignoring every preference after the first.

= Non-LNH methods =

Nearly all voting methods not discussed above fail LNH, including score voting, highest medians, Borda count, and all Condorcet methods. The Condorcet criterion is incompatible with later-no-harm (assuming the resolvability criterion, i.e. any tie can be removed by a single voter changing their rating).Douglas Woodall (1997): [http://scorevoting.net/Woodall97.pdf Monotonicity of Single-Seat Election Rules], Theorem 2 (b)

Bloc voting, which allows a voter to select multiple candidates, does not satisfy later-no-harm when used to fill two or more seats in a single district, although the single non-transferable vote does.

Examples

= Anti-plurality =

{{Main|Anti-plurality voting}}

Anti-plurality elects the candidate the fewest voters rank last when submitting a complete ranking of the candidates.

Later-No-Harm can be considered not applicable to Anti-Plurality if the method is assumed to not accept truncated preference listings from the voter. On the other hand, Later-No-Harm can be applied to Anti-Plurality if the method is assumed to apportion the last place vote among unlisted candidates equally, as shown in the example below.

class="wikitable collapsible collapsed"

!Examples

; Truncated Ballot Profile

Assume four voters (marked bold) submit a truncated preference listing A > B = C by apportioning the possible orderings for B and C equally. Each vote is counted \tfrac{1}{2} A > B > C, and \tfrac{1}{2} A > C > B:

{| class="wikitable"

! # of voters !! Preferences

2A ( > B > C)
2A ( > C > B)
1B > A > C
1B > C > A
1C > A > B
1C > B > A

Result: A is listed last on 2 ballots; B is listed last on 3 ballots; C is listed last on 3 ballots. A is listed last on the least ballots. A wins.

; Adding Later Preferences

Now assume that the four voters supporting A (marked bold) add later preference C, as follows:

class="wikitable"

! # of voters !! Preferences

4A > C > B
1B > A > C
1B > C > A
1C > A > B
1C > B > A

Result: A is listed last on 2 ballots; B is listed last on 5 ballots; C is listed last on 1 ballot. C is listed last on the least ballots. C wins. A loses.

;Conclusion

The four voters supporting A decrease the probability of A winning by adding later preference C to their ballot, changing A from the winner to a loser. Thus, Anti-plurality doesn't satisfy the Later-no-harm criterion when truncated ballots are considered to apportion the last place vote amongst unlisted candidates equally.

|}

= Borda count =

{{Main|Borda count}}

class="wikitable collapsible collapsed"

!Examples

This example shows that the Borda count violates the Later-no-harm criterion. Assume three candidates A, B and C and 5 voters with the following preferences:

{| class="wikitable"

! # of voters !! Preferences

3A > B > C
2B > C > A

; Express later preferences

Assume that all preferences are expressed on the ballots.

The positions of the candidates and computation of the Borda points can be tabulated as follows:

class="wikitable"

! candidate !! #1. !! #2. !! #last !! computation !! Borda points

A3023*2 + 0*1bgcolor=#ffbbbb|6
B2302*2 + 3*1bgcolor=#bbffbb|7
C0230*2 + 2*1bgcolor=#ffbbbb|2

Result: B wins with 7 Borda points.

;Hide later preferences

Assume now that the three voters supporting A (marked bold) would not express their later preferences on the ballots:

class="wikitable"

! # of voters !! Preferences

3A
2B > C > A

The positions of the candidates and computation of the Borda points can be tabulated as follows:

class="wikitable"

! candidate !! #1. !! #2. !! #last !! computation !! Borda points

A3023*2 + 0*1bgcolor=#bbffbb|6
B2032*2 + 0*1bgcolor=#ffbbbb|4
C0230*2 + 2*1bgcolor=#ffbbbb|2

Result: A wins with 6 Borda points.

;Conclusion

By hiding their later preferences about B, the three voters could change their first preference A from loser to winner. Thus, the Borda count doesn't satisfy the Later-no-harm criterion.

|}

= Copeland =

{{Main|Copeland's method}}

class="wikitable collapsible collapsed"

!Examples

This example shows that Copeland's method violates the Later-no-harm criterion. Assume four candidates A, B, C and D with 4 potential voters and the following preferences:

{| class="wikitable"

! # of voters !! Preferences

2A > B > C > D
1B > C > A > D
1D > C > B > A

;Express later preferences

Assume that all preferences are expressed on the ballots.

The results would be tabulated as follows:

class=wikitable

|+ Pairwise election results

colspan=2 rowspan=2 |

| colspan=4 bgcolor="#c0c0ff" align=center | X

bgcolor="#c0c0ff" | A

| bgcolor="#c0c0ff" | B

| bgcolor="#c0c0ff" | C

| bgcolor="#c0c0ff" | D

bgcolor="#ffc0c0" rowspan=4 | Y

| bgcolor="#ffc0c0" | A

|

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#ffe0e0" | [X] 1
[Y] 3

bgcolor="#ffc0c0" | B

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

|

| bgcolor="#ffe0e0" | [X] 1
[Y] 3

| bgcolor="#ffe0e0" | [X] 1
[Y] 3

bgcolor="#ffc0c0" | C

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#e0e0ff" | [X] 3
[Y] 1

|

| bgcolor="#ffe0e0" | [X] 1
[Y] 3

bgcolor="#ffc0c0" | D

| bgcolor="#e0e0ff" | [X] 3
[Y] 1

| bgcolor="#e0e0ff" | [X] 3
[Y] 1

| bgcolor="#e0e0ff" | [X] 3
[Y] 1

|

colspan=2 bgcolor="#c0c0ff" | Pairwise election results (won-tied-lost):

| bgcolor=#ffbbbb|1-2-0

| bgcolor=#bbffbb|2-1-0

| bgcolor=#ffbbbb|1-1-1

| bgcolor=#ffbbbb|0-0-3

Result: B has two wins and no defeat, A has only one win and no defeat. Thus, B is elected Copeland winner.

;Hide later preferences

Assume now, that the two voters supporting A (marked bold) would not express their later preferences on the ballots:

class="wikitable"

! # of voters !! Preferences

2A
1B > C > A > D
1D > C > B > A

The results would be tabulated as follows:

class=wikitable

|+ Pairwise election results

colspan=2 rowspan=2 |

| colspan=4 bgcolor="#c0c0ff" align=center | X

bgcolor="#c0c0ff" | A

| bgcolor="#c0c0ff" | B

| bgcolor="#c0c0ff" | C

| bgcolor="#c0c0ff" | D

bgcolor="#ffc0c0" rowspan=4 | Y

| bgcolor="#ffc0c0" | A

|

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#ffe0e0" | [X] 1
[Y] 3

bgcolor="#ffc0c0" | B

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

|

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

bgcolor="#ffc0c0" | C

| bgcolor="#c0c0c0" | [X] 2
[Y] 2

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

|

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

bgcolor="#ffc0c0" | D

| bgcolor="#e0e0ff" | [X] 3
[Y] 1

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

| bgcolor="#c0c0c0" | [X] 1
[Y] 1

|

colspan=2 bgcolor="#c0c0ff" | Pairwise election results (won-tied-lost):

| bgcolor=#bbffbb|1-2-0

| bgcolor=#ffbbbb|0-3-0

| bgcolor=#ffbbbb|0-3-0

| bgcolor=#ffbbbb|0-2-1

Result: A has one win and no defeat, B has no win and no defeat. Thus, A is elected Copeland winner.

;Conclusion

By hiding their later preferences, the two voters could change their first preference A from loser to winner. Thus, Copeland's method doesn't satisfy the Later-no-harm criterion.

|}

= Schulze method =

{{Main|Schulze method}}

class="wikitable collapsible collapsed"

!Examples

This example shows that the Schulze method doesn't satisfy the Later-no-harm criterion. Assume three candidates A, B and C and 16 voters with the following preferences:

{| class="wikitable"

! # of voters !! Preferences

3A > B > C
1A = B > C
2A = C > B
3B > A > C
1B > A = C
1B > C > A
4C > A = B
1C > B > A

; Express later preferences

Assume that all preferences are expressed on the ballots.

The pairwise preferences would be tabulated as follows:

class="wikitable" style="text-align:center"

|+ Matrix of pairwise preferences

d[*,A]d[*,B]d[*,C]
d[A,*]

| || bgcolor=#ffdddd|5 || bgcolor=#ddffdd|7

d[B,*]

| bgcolor=#ddffdd|6 || || bgcolor=#ddffdd|9

d[C,*]

| bgcolor=#ffdddd|6 || bgcolor=#ffdddd|7 ||

Result: B is Condorcet winner and thus, the Schulze method will elect B.

== Hide later preferences ==

Assume now that the three voters supporting A (marked bold) would not express their later preferences on the ballots:

class="wikitable"

! # of voters !! Preferences

3A
1A = B > C
2A = C > B
3B > A > C
1B > A = C
1B > C > A
4C > A = B
1C > B > A

The pairwise preferences would be tabulated as follows:

class="wikitable" style="text-align:center"

|+ Matrix of pairwise preferences

d[*,A]d[*,B]d[*,C]
d[A,*]

| || bgcolor=#ffdddd|5 || bgcolor=#ddffdd|7

d[B,*]

| bgcolor=#ddffdd|6 || || bgcolor=#ffdddd|6

d[C,*]

| bgcolor=#ffdddd|6 || bgcolor=#ddffdd|7 ||

Now, the strongest paths have to be identified, e.g. the path A > C > B is stronger than the direct path A > B (which is nullified, since it is a loss for A).

class="wikitable" style="text-align:center"

|+ Strengths of the strongest paths

p[*,A]p[*,B]p[*,C]
p[A,*]

| || bgcolor=#ddffdd|7 || bgcolor=#ddffdd|7

p[B,*]

| bgcolor=#ffdddd|6 || || bgcolor=#ffdddd|6

p[C,*]

| bgcolor=#ffdddd|6 || bgcolor=#ddffdd|7 ||

Result: The full ranking is A > C > B. Thus, A is elected Schulze winner.

; Conclusion

By hiding their later preferences about B and C, the three voters could change their first preference A from loser to winner. Thus, the Schulze method doesn't satisfy the Later-no-harm criterion.

|}

Criticism

Douglas Woodall writes:

{{blockquote|[U]nder STV the later preferences on a ballot are not even considered until the fates of all candidates of earlier preference have been decided. Thus a voter can be certain that adding extra preferences to his or her preference listing can neither help nor harm any candidate already listed. Supporters of STV usually regard this as a very important property, although it has to be said that not everyone agrees; the property has been described (by Michael Dummett, in a letter to Robert Newland) as "quite unreasonable", and (by an anonymous referee) as "unpalatable".Woodall, Douglas, Properties of Preferential Election Rules, [http://www.votingmatters.org.uk/ISSUE3/P5.HTM Voting matters - Issue 3, December 1994]

}}

See also

Notes

{{reflist|group=note}}

Bibliography

  • D R Woodall, "Properties of Preferential Election Rules", Voting matters, Issue 3, December 1994 [http://www.votingmatters.org.uk/ISSUE3/P5.HTM]
  • Tony Anderson Solgard and Paul Landskroener, Bench and Bar of Minnesota, Vol 59, No 9, October 2002. [https://web.archive.org/web/20090204174848/http://www2.mnbar.org/benchandbar/2002/oct02/voting.htm]
  • [http://rangevoting.org/BrownVsmallwood.pdf John Brown, Jr. v. W. H. Smallwood], 130 Minn. 492, 153 N.W. 953 (July 30, 1915)

References

{{reflist}}

{{voting systems}}

Category:Electoral system criteria