What: Claim that IRV is the only non-monotonic election method
IRV FactCheck Rating: FALSE
Terry Reilly has often faulted IRV for being non-monotonic. It is one of his favorite complaints to make about IRV.
IRV is not monotonic because in certain situations giving a candidate greater preference, for example marking the candidate as your first choice instead of your second choice, can make that candidate lose. Likewise, in certain situations giving a candidate less preference, for example marking the candidate as your second choice instead of your first choice and raising another candidate to first choice, can in certain situations make the downgraded candidate win.
A perfect election method would be monotonic. So saying that IRV is not monotonic is really just another way of saying IRV is not perfect. But no election method is perfect.
However when talking about the problem, Terry Reilly is playing a game of Your Epidermis is Showing. He takes commonly accepted election method behavior and exploits people's unfamiliarity with voting theory in order to skew their judgement. Terry Reilly knows that if he gives a quick example only in relation to IRV, many people will erroneously jump to the conclusion that being non-monotonic is an intolerable problem with IRV and they are better off keeping or returning to traditional election methods.
What Terry Reilly doesn't tell people is that traditional two-election runoff systems are also non-monotonic. In fact, some of the examples of non-monotonicity that he gives for IRV work just as well with a traditional two-election runoff system, the kind of voting that is the most popular alternative to IRV. If Terry Reilly really thought that being non-monotonic was such a serious flaw in an election method, you have to wonder why he would celebrate attempts to return to traditional two-election runoffs (both attempts failed).
If you have voted in a local election where the top two candidates advance to a later runoff election, you've used an election method that is non-monotonic and lived to tell about it. If you have voted in a partisan primary for state or federal office where the winner was chosen from two or more candidates in the November general election, you have likely confronted bigger challenges to having your voice heard than the potential problems of deciding the contest with non-monotonic elections.
But if you have ever enjoyed the benefits of voting in an IRV contest or would consider doing so, Terry Reilly wants to warn you: your epidermis is showing.
You might not have expected such a misleading and exploitative argument from someone who promotes himself as a former chair of the Campaign Finance Review and and Ethics Board for San Jose, California. Unfortunately, sometimes Terry Reilly goes even further by explicitly claiming that IRV is the only election method that is not monotonic.
In a recent email and news article, Terry Reilly did not just feature his criticism that IRV, also known as ranked choice voting (RCV), is non-monotonic. He also claims that being non-monotonic “is unique to the RCV style voting”.
So Terry Reilly's claim that only IRV is non-monotonic earns an IRV FactCheck rating of FALSE.
In the same email and news article Terry Reilly has also misleadingly expanded on IRV being non-monotonic with the vague and conceptually confused claim that “your vote can hurt your candidate rather than help them.” To be clear, marking your favorite candidate as your first choice never hurts that candidate's chances of winning compared to you not voting at all. Marking a second-choice candidate will never keep your first-choice candidate from winning compared to not marking any second-choice candidate. But like all other election methods, with IRV there are some situations where voting strategically can give a better election result for some voters. An important advantage of IRV is that the opportunities for strategic voting tend to be less than with traditional election methods or many other alternatives. But Terry Reilly counts on many people reading more into his vague generalization than is justified.