In the Dec. 10 Letter to the Editor “Taxation without representation?“, the author argued that Lexington’s Town Meeting Members (TMMs) did not truly represent their constituents in the recent Debt Exclusion vote. The argument was based on a comparison between the yes/no vote totals from voters in different precincts and the corresponding yes/no votes cast by their TMMs.
The core of the argument is that while 91% of Town Meeting Members voted in favor of the measure, only 62% of the town’s voters did so. The author presents this discrepancy as evidence that, overall, TMMs failed to accurately represent voter sentiment on the issue.
However, I would argue that the 62% vs. 91% discrepancy is exactly what one would expect from a district-based representative democracy. Consider the following simplified scenario:
- Each TMM represents 45 voters in the neighborhood where they reside and faithfully reflects the sentiment of those voters.
- If at least 23 of the 45 voters support the measure, the TMM votes “yes”; otherwise, the TMM votes “no.”
- Overall voter sentiment across the town is 62% in favor and 38% opposed.
We can now calculate the probability that a given TMM would vote “yes” under this simplified model. This is a straightforward binomial calculation, illustrated in the distribution below:
The probability that at least 23 out of 45 voters support the measure, leading the TMM to vote “yes”, is:
P(Bin(45, 0.62) >= 23) = 0.9498
In other words, even with an overall voter support level of 62%, we would expect roughly 95% of TMMs to vote “yes.” Seen in this light, the observed 91% “yes” vote among TMMs aligns very closely with what the underlying voter sentiment would predict.
Of course, this is only a simplified model. In reality, voter preferences are not independent, precincts differ in composition, and TMMs exercise judgment informed by discussion and debate. That said, this simple analysis shows that the observed discrepancy is not evidence of misrepresentation. On the contrary, it is entirely consistent with, and arguably supportive of, the conclusion that TMMs, as a body, represent voter sentiment quite well within a district-based representative system.
Best regards,
Dazhi
It seems our analyses rest on different assumptions. First we need to ask: What is the role of a TMM? Are they functioning like an Electoral College (your assumption, effectively winner-takes-all), or as a representative sub-set of voters (my assumption)?
If TMMs were acting like an Electoral College, we would expect zero ‘No’ votes. In reality, about 10% of TMM votes were ‘No’—not zero—despite every precinct majority voting ‘Yes.’ So the ‘district-based representative democracy’ model isn’t supported by the data.
If TMMs are meant to be a representative sub-set, then some deviation from the public vote is normal, but the size matters. The data show that over 30% of TMM votes would need to be reassigned to match constituent preferences. That’s a large effect, not a minor variation.
Overall, the data suggest TMMs are voting based on their own preferences (and who chooses to run), rather than truly representing their precincts.
So the real question is: What should we do about it?
Hi Lana, the model presented here is not an Electoral College–style, winner-takes-all system, which, as you correctly noted, would indeed produce a 100% “yes” vote at town meeting.
Instead, this is better described as a “majority-rule” model, where each TMM casts a vote based on the simple majority of their constituents.
Importantly, as the number of TMMs increases, and the number of voters each represents decreases, the aggregate TMM voting pattern will increasingly approximate the underlying voter distribution. For example, if we were to double the number of TMMs so that each represents about 23 voters on average, the expected “yes” vote among TMMs would drop from roughly 95% to about 88%, even if overall voter sentiment remains unchanged.
As others have noted below your post, pure proportional representation would be difficult to implement within Lexington’s current nonpartisan system. Moreover, I don’t believe Lexington voters consistently decide local issues strictly along party lines.
As discussed earlier, increasing the number of TMMs would help reduce the representational distortion. More thoughtful strategies around TMM candidacy, particularly with respect to key issues, could also improve alignment with voter sentiment. I’m less certain about the impact of ranked-choice voting in this context, as it would require additional modeling to understand its effects.
Another option just came to mind: we could allocate a portion of TMM seats—say, 30%—as at-large positions. This would allow voters with minority viewpoints across precincts to aggregate their support town-wide, increasing the likelihood that those perspectives are represented at town meeting, while still preserving strong neighborhood-based representation for the remaining seats.
Or, we could restrain the number of terms served to 2 – across all governing positions. And place the new candidates names upfront in the ballot, followed by the “old ones” … that will get fresh blood flowing and keep it aligned with constituents… and maybe have a rule-book for the TMMs, with rights, responsibilities, and transparency expectations.
I am moving onto possible solutions 🙂
And we should move the Town election from a Monday in March, when 85% of voters stay home (and most Town-wide positions are not even contested) to the 2nd Tuesday in November when most people expect an election to be held.
The small group of perpetual incumbents (like Joe Pato, who was “elected” 4 times without a single opponent to the Select Board, and re-elected this past March to 1 of 2 open seats with 4 candidates for those 2 seats) has consistently opposed moving our Town election from March to November since I first proposed this two decades ago. Why? I guess because it’s nice to get “elected” and re-“elected” without competition by the small group of insiders that the 15% of registered voters constitutes.