City of Hudson's weighed voting system under scrutiny

[jimrtex]

And part 2

You must be logged in to read this quote.

I don't believe that there is a mathematical formula.

Power share does not vary proportionately with voting weight.   As we saw above, we could adjust the weights from (20, 20, 20, 20, 20) to (30, 20, 20, 20, 10) and see no change in voting power, and then suddenly at (31, 20, 20, 20, 9) there was a catastrophic change.

On the other hand, I've noted that by using simple population-based weights, but changing the threshold for success to  something slightly different than 5%+1, that I could get just as good proportionality between population share and voting power, as by changing the voting weights.   I sort of did this in my examples above.  I initially had 100 total votes, so that a majority was 51 or 51%.  In my last example, I used 101 total votes, and a majority of 51, or 50.5%.  And in an unweighted 5-member body, a majority is 3/5 or 60%.

Simple-minded judges won't accept a redefinition of "majority" even if the new percentage is only slightly different and still represents the notion of "bare concurrence:.

But we can trick the judges by adding an extra weighted vote.  Let's say that the at-large member has a 10% voting share.   Then if he joins with district members who have votes representing more than 40% of the total vote, then the motion passes.  A district member who is critical to getting past 40%, is also critical to the combination including the at-large member.

But we could boost the at-large member to 11%, and proportionately reduce the other shares.  The district members now only produce more than 39% of the total.   In the original version they needed 40/90 of the district vote (or 44.4%).  In the second they needed 39/89 (or 43.8%).

We've subtly switched from "majority" which is fixed at 50%+, to "significant district concurrence" which can float.

This currently happens in Hudson, where the President's voting weight varies quite widely between votes for a simple majority, and the 2/3 and 3/4 super-majorities.

Use of an odd number of members, plus a president may have other problems.  Consider an unweighted council.  If the district members vote 2:3, the president can not change the result.  Only if they vote 3:2 can he in effect veto their decision.

If there were an even number of members, he has a casting vote, whether formally or informally.  If the district members vote 3:3 he breaks the tie.  If the district members vote 4:2, then he has no effect.

You must be logged in to read this quote.

I use Computer Algorithms for Voting Power Analysis, in particular the program lpgenf.

If you set your voting weights in a column, you can copy and paste them directly in the "Weights" box, and be sure to set the correct "quota" or majority.  You can then copy and paste the results from the web page into your spreadsheet using the "Match Destination Formatting" which will simply grab the numbers.

You are most interested in "Swings".   The Normalized Banzhaf Index is the power share among the voting members.   But in Hudson you want to use the power share among the district members, which is each member's swings divided by the total number of swings, excluding those of the president.

Add in a column with population, you can calculate population share and power share, and deviation (power-share / population-share - 1).   All of these are expressed as percentages.  I usually copy the results from the program over to the right side of my spreadsheet, and then copy the swings into a column used for computation.

The spreadsheet that I use is set up specifically for Hudson (11-members).  It brute force generates the 2048 voting combinations, and then counts critical swings, etc.

In either case, using the program or the spreadsheet, iteration is doable, but not automatic.  I could clean up the spreadsheet and send you a copy.

You must be logged in to read this quote.

The 10% limit may not be applicable to local governments.  See Abate v Mundt and
Roxbury Taxpayers Alliance v Delaware County Board of Supervisors  In the latter case, the deviation range was 16.79%.

5% of 1280 is 26, which doesn't permit much flexibility when many individual blocks have more than 26 persons.

Nearly equal-population wards can be a problem with weighted voting.   See Cortland County, which has a 17-member legislature where the districts are either combinations of towns, or divisions of the city of Cortland or the town of Cortland.   Each member has a voting weight equivalent to the population they represent.   But the districts are so similar in size, that any combination of 9 members represents a majority of the population (and therefore a majority of the legislative votes), and no combination of 8 members represents a majority.

[Torie]

Thanks Jim. Again, I, and all of the citizens of Hudson,  are very much in your debt.

[jimrtex]

This is the beginning of an analysis of the problems with weighted voting when there are only a few number of members.  I am using the official Hudson numbers, so as to not dirty up the analysis with other issues.   Each ward was assigned a weight proportional to its population, such that they added up to 1000.  Thus each weighted vote unit is equivalent to 0.1% of the total population of 6.403 persons.  A simple majority is 50.1%.

W  Pop Vote
1  770 120
2 1281 200
3 1142 179
4  725 113
5 2485 388

With five members, there are only 32 (2⁵) combinations, and at most 32 possible vote totals.

0
113
120
179
200
233
292
299
313
320
=== 1/3
379
388
412
433
492
499
=== 1/2
501
508
567
588
612
621
=== 2/3
680
687
701
708
767
800
821
880
887
1000

If we consider any vote that receives more than 2/3 support, or less than 1/3 support as representing a consensus for or against, then there are only 12 combinations that are close or somewhat contentious (for Hudson, this is not really true, since with Ward 5 having almost 39% of the population, their support could be critical to passage of motions that receive well in excess of 2/3 support.

If we look at the actual critical votes, there are only 22 of them, which means that each critical swing represents 4.5% of the total, and that it would extremely difficult to match population share (4.5% of 20% is 22.5%, which illustrates the extreme coarseness of the critical votes)

The actual voting power is:

Ward            Vote   Swing  R.Pop.  R.Pow   Dev. 
Ward 1           120       2  12.03%   9.09% -24.40%
Ward 2           200       2  20.01%   9.09% -54.56%
Ward 3           179       2  17.84%   9.09% -49.03%
Ward 4           113       2  11.32%   9.09% -19.71%
Ward 5           388      14  38.81%  63.64%  63.97%

It turns out that Ward 5 has an extreme share of the power and the remaining four wards have equal power.   It turns out that Ward 5 and any other ward represent a majority of the population (Ward 5 + Ward 4 = 50.1%, Ward 5 + Ward 2 = 58.8%).  Further, the three largest of the wards other than Ward 5, do no represent a majority.   To overcome Ward 5 opposition, all four other wards must vote together.

For the 4 wards other than Ward 5, their critical votes are (1) When they join with Ward 5; or (2) When they join with the three other smaller wards.

For Ward 5, the 14 critical votes are when they join with any one of the 4 smaller wards; with any of the 6 possible pairs of the four smaller wards; or with any of 4 triplets of the smaller wards (14 = 4 + 6 + 4).

The total range of the deviations is 118.5%.   So while we can get quite close to matching population share,

Ward            Vote   Vote   R.Pop.  R.Wgt   Dev. 
Ward 1           120     120  12.03%  12.00%  -0.21%
Ward 2           200     200  20.01%  20.00%  -0.03%
Ward 3           179     179  17.84%  17.90%   0.36%
Ward 4           113     113  11.32%  11.30%  -0.20%
Ward 5           388     388  38.81%  38.80%  -0.03%

Total deviation 0.57%, we are far far away on voting power.  A compromise is likely to give a bad mismatch of voting weight to population AND votiing power to population.

[jimrtex]

On the Subject of the Weighted Vote purports to show the equal population map used for the 2003 referendum.

I don't think that is the map that was voted on.  The resolution passed in spring 2003 match this map:



This map is more plausible since it addresses the issue of underpopulation of wards 1 and 4 in a straightforward way, shifting them eastward along Warren, and then moving the boundary between 3 and 5 northward.   The other map combined Wards 1 and 4 catty-corner-wise.

Both have a problem of Ward 3 including the prison population, which was larger in 2000 and 2010.   Note the boundary includes the driveway off of Harry Howard in front of the Firemen's Home.  This is a misapprehension of census geography.  Census block boundaries use visible features such as roads.  They aren't loops to capture particular building.  See the cemetery, which has many census blocks, though with no people living there.

Looking closer at the map, it is actually a 6-district plan.

Adding the ward populations on the map, they total 6827.

The 2000 census population was 7524.
The 2010 census population was 6713.
The 2010 population after prison adjustments was 6403.

But the 2000 census population for Ward 1 (as shown on their map) is correct.

The difference between 7524 (2000 census) and 6827 (as shown on map) is 697, which happens to be the 2000 population of the census block containing the prison.   In 2000, the RR tracks were used as a census block boundary which did a better job of separating the prison from the persons along the south side of Union, but not entirely.

The population for Wards 1, 2, 3, and 5 appear to match the 2000 census, if we assume that the population of the prison block was entirely in the prison, which it was not.  Those are clearly houses on the south side of Union, east of where the RR crosses, as well as west of Worth Avenue.   I can't quite get 4 and 6 to match up.

To get the correct population for Wards 3 and 5, the parrot's beak has to be included in Ward 3.

But unless they did a switch on the resolutions, this map was not the one on the ballot.

[jimrtex]

Hofstra Report on Hudson governance

[jimrtex]

Some visualization is in order.

This is based on the current 5 wards, each electing one alderman.   The voting strength of each alderman is proportional to the population of their ward.   The weights are calculated such that they sum to 1000, so that one unit corresponds to 0.1% of the population (6.403 persons).

For example, the alderman from Ward 1, which has 12.03% of the population has 120 votes. Normalization makes it easy to convert to percentages.   For example, a majority of 501 votes is 50.1% of the total votes.

I am using the erroneous populations used to calculate the current weights because it would confuse matters to switch to the correct populations, and might lead to the conclusion that a set of particular populations causes the problem.   The problem is more fundamental.  Weighted voting does not work well, if at all, with small bodies.

Consider the case of a 3-member body.   Under unweighted voting, any combination of two members prevails on any vote.   To change this dynamic via weighted voting, we would have to create situations where a single member could prevail, or alternatively where all 3 members would be required.

But if Adams voting for a measure causes it to prevail, and Brown and Cook can not prevail, then Adams is a dictator, and Brown and Cook are mere observers (example Adams 51, Brown 28, Cook 21).

If the votes of the three members are weighted such that it requires 2 members to prevail (for example: Adams 44, Brown 32, Cook 24; Adams+Brown:Cook = 76:24;  Adams+Cook:Brown = 68:32; Brown+Cook:Adams 56:44; then the dynamics are identical to an unweighted system where each member had one vote.  If the weights were based on population, you simply have a system of grossly non-equipopulous districts, where Adams' district has an excess population of 32%, and Cook's district is 28% underpopulated.  The weights in this case would be tissue paper attempting to hide a OMOV violation.

In a 5-member body it is slightly better.   In an unweighted system, any combination of 3 members will prevail.   To change the dynamics, you have to make it possible for some combinations of two members to prevail (which also mean that the other 3 members can not block the action, or prevail in their own right).   While this is possible, there is a risk that the two most powerful members could become too powerful.  If the two can get together and make deals, they would be able to win every vote.

If the weights are such that only 3-member combinations can prevail, and all 3-member combinations prevail, then again you are back to non-equipopulous districts with equal voting power.   This can also happen in a larger body. Cortland County has a 17-member legislature, in which any 9-members can prevail, even those from the 9 smallest districts, yet each member has a weighted vote (the same as their district's population).  Cortland County could give each member one vote and a glass of water, and the results would be the same (save if the water led to clearer thinking).

The following table shows the voting power of the 5 ward alderman, if they had a voting weight proportional to their population.



Ward 1 has a population of 770, which is 12.03% of the total.  There voting weight is 120, which is 12.00% of the total.  There is, as we would expect, extremely good conformance between population and voting weight.  The small relative error (0.21%) is due to rounding.

There would be two voting combinations where the alderman from Ward 1 would be critical. The combination would prevail, but if he withdrew his support, it would fail.   His vote was necessary for the combination to prevail.   In the other 14 possible combinations in which the alderman voted Aue, either the combination would not have a majority and fail, or his support would be superfluous.  For example, if the vote was unanimous, it would not matter that the alderman from Ward 1 switched to No.  It would still have 88%.

There are 22 critical votes among all 5 alderman.  The Ward 1 alderman's share would be 2/22 or 9.09%.  This is his relative Banzhaf index, or voting power share.  Comparing 9.09% his voting power share, to the 12.02% population share, results in a -24.40% deviation.

Overall among the five aldermen, the range of deviations is -54.56% to 63.97%, for a total range of 118.53%, which is extremely outside the generally accepted limits of 10% for local governments.  The standard deviation is 42.55%, and RMS is 52.23%, which are quite poor (we would expect something around 3% or so).   The RMS takes into account the population of the wards.   In this case, the error is largest wards is the greatest, so a relatively large share of the voters is excessively advantaged (Ward 5) or excessively disadvantaged (Wards 2 and 3).

The small number of critical votes (22) makes it an extremely coarse measurement of power.  If the alderman from Ward 1 increased his critical votes from 2 to 3, his power share would increase from 9.09% to 13.63%.   The relative deviation would go from a deficit of -24.40%, to an excess of 13.39%.   Closer, but not close.  It would like measuring someone who is 5'7" as either 5 foot of 6 foot.   6 foot is the better answer, but not a very good answer.

This coarseness is directly related to the small number of members.  Even if all members had an equal vote, the number of critical combinations is 30 (I suspect this is a theoretical maximum for a five member body).   Using the current weights for the 11-member Common Council, there are 2224 critical votes (this ignores the quite dubious assumption that the aldermen from a ward will vote opposite each other half the time, and that a voter from a ward has power when his two aldermen are cancelling each other out, and loses power if they were to agree).

It may also be quite difficult to improve the voting power for one alderman without making it worse for others.  If you push in one place, it bulges out somewhere else.   Dr. Papayanopoulos literally finds possible voting weights by trial and error, having a computer test nearly a million different sets of voting weights, before spitting out a few favored plans.

The following chart illustrates how the voting power is calculated (It is helpful to see this at a larger scale.   In Firefox, right click, and the click on View Image)



Each row of circles represents the votes of an alderman (Ward 5 on the top and so on).   Each column represents the votes for a particular combination of alderman, with the horizontal placement representing the total votes for that combination.   An open circle indicates a No vote.  A filled blue circle represents an Aye vote.  Those on the left side are votes that failed to secure a majority.  Those on the right side receive a majority, but the vote was not critical to success.  A red filled circle is a critical vote, where if the vote were switched to No, the combination would no longer prevail.

At the extreme left is the case where a motion received zero votes because everyone voted against it.  At the extreme right is a unanimous vote, which results in a 1000:0 victory.  No member's vote is critical since any motion with 4 members will still prevail.   A combination of wards 3 and 5 results in a prevailing vote of 567:433.  Both ward 3 and 5 are critical, since the removal of either would cause the motion to fail.  A combination of wards 1, 4, and 5 results in a prevailing vote of 621:379, but only Ward 5 is critical.  Removal of either Ward 1 or Ward 5 does not cause the motion to fail.

If we count the filled red circles, we will find 14 critical votes for Ward 5, and 2 each for other 4 wards, just as is shown in the above table.   Carefully looking at the different combinations, we see that Ward 1 will be critical if it combines with Ward 5 or if it combines with all 4 smaller wards.   The same is true for Ward 2, 3, and 4.   Ward 5 and any other ward represents a majority of the population, a bare majority of 50.1% for wards 4 and 5, and a comfortable margin of 58.8% for wards 2 and 5.   But since passage of a motion is a TRUE/FALSE proposition, there is effectively no difference between the two.  Wards 4 and Wards 2 effectively have the same power, despite Ward 2 having a 76% larger population.

Ward 5 is given credit for a critical vote for combinations involving itself and Ward 1; or itself and Ward 4; or itself and both Wards 1 and 4.  In a sense this is double counting.   Imagine the freshman alderman wants to get a stop sign in his ward.   Being new, he goes to the other members, and starts with Ward 1.   The alderman from Ward 1 agrees, in exchange for support for a paving project in his ward.   But they still are short votes.   The Ward 4 alderman suggest going to Ward 2 next, but the more experience alderman from Ward 1, knows to go to Ward 5.  The alderman from Ward 5 may ask for sodding a park, plus another project.

If the alderman from Ward 5 made a deal with each of the other alderman, then there would be 4 projects in Ward 5, and 1 in each of the other wards.   Ward 5 would get 50% of the projects, despite having only 39% of the population.  Wards 1 and Wards 4 would get about a proportional share, and ward 2 and 3 would be shortchanged.

[jimrtex]

This is Torie's proposed 5-ward map, which will apparently be subject to a referendum in the November 2015 election, either by the Common Council or the initiative.



It keeps the current split of the block off of Harry Howard.   Since that split was the apparent source of misadminstration of elections, and not taken into account when calculating the voting weights, I don't think it should be perpetuated.   The only reason to use it, is to get better population equality.  But you would not split a block to get better population equality, and if yu did you wouldn't run the dividing line through somebody's living room.

This alternative eliminates the split block.



Torie's proposal uses Rope Alley as a boundary between wards 2 and 4.   While alleys are named in Hudson, and the census bureau uses them as block boundaries, they are still alleys, lined by garages, with the houses at the front of the lots.

This alternative switches the split between wards 2 and 4 to be between Warren and Columbia.



This leaves the area along Mill Street east of 2nd street isolated from the rest of the 4th ward.  The only reason they would be included is because of the census block does not recognized settlement patterns on a city scale.   The block is currently split betweenwards 2 and 4.

This alternative restores the area to the 2nd ward.  The split is different than that of the Harry Howard block, because there are no houses near the boundary, and the area in Ward 2 only has a couple of dozen persons.

[Torie]

Here is a little Hudson update. The smell of litigation is in the air. Of course this is but the tip of the iceberg, but this one is well, an exemplar of upstate politics in New York. It's a contact sport.   Mr. Nastke is the Pub Elections Board member by the way, and Virginia Martin is the Dem member. It's one of the reasons I changed party registration. The Dems around my hood tend to be - well - more law abiding. Who knew?

My comment generated some interesting emails from one of the power players, but whatever.

[muon2]

Are those the only two election board members? What official role do they have in NY?

[Torie]

Are those the only two election board members? What official role do they have in NY?

As I understand it, in Columbia County, there is a Pub Commissioner and a Dem one, and they set policy, and must agree to do anything. One of their responsibilities, is to register voters and maintain the voter rolls. So they review the Ward lines, and maintain a set or range of addresses within each ward, and when folks register, they cross check those addresses against the voter's address, and passing the voter to a ward (and within the 5th ward, to one of the two "voting districts."

In Hudson, we had ward maps that were wrong, so the population counts were wrong, but the population count does not involve the Board of Elections. However, for the reasons stated, and for unknown reasons, it turns out that while a map error caused the mis-assignment of voters in the Columbia triangle, as well as a population error, voters along Harry Howard Way while counted correctly, were in the wrong ward on the voting rolls. Other voters in the area were both counted correctly, and are voting in the right ward, even though the map shows the voters in the wrong ward, as if the map was ignored. And then still other voters who are in the right ward on the voting rolls, but were counted incorrectly. Anyway, what the Board of Elections needs to do is get the voting rolls right, with voters assigned to the correct ward.

[muon2]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

[Torie]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

Yes, they have staff. Other than the Columbia triangle, most of the errors occurred with respect to the 2010 census. At about that time, the city clerk used the erroneous BOE map rather than the city charter to count ward populations (and did some block allocation errors), which map the BOE said was just an approximate guide and not exact (somebody is hiding something),  and Crosswinds was built, nobody is explaining why the voters were put in the wrong ward, other than three houses on that side of the street were in the wrong ward, and apparently had been for years for unknown reasons, and nobody bothered to check either the city charter or the erroneous map.  So either the BOE are sort of incompetent drones, or somebody was playing ball with what certain politicians wanted (and still want). We will never know probably.

[jimrtex]

Here is a little Hudson update. The smell of litigation is in the air. Of course this is but the tip of the iceberg, but this one is well, an exemplar of upstate politics in New York. It's a contact sport.   Mr. Nastke is the Pub Elections Board member by the way, and Virginia Martin is the Dem member. It's one of the reasons I changed party registration. The Dems around my hood tend to be - well - more law abiding. Who knew?

My comment generated some interesting emails from one of the power players, but whatever.

Under New York Statute, Election Code 4-100, the legislative body of a city or town has responsibility for defining election districts ("election district" is New York-ese for voting precincts).  There is an exception if a city or town has requested the Board of Elections (BOE) to define the election districts.  Read carefully the last sentence of the section, where it is absolutely clear that the legislative body (ie Common Council) has primary responsibility.

Even if the City of Hudson has in the past has made a request to the BOE, surely they can rescind that request.

It is both a legal and functional requirement that election districts not cross ward boundaries.  Surely the Board of Elections is not going to claim that they have the authority to override the charter.  Even the Common Council is unable to do that.  They may only make a proposal that would take effect upon approval by the voters.

The Board of Elections might have authority to define the internal boundary between 5-1 and 5-2, but it appears that this does not comply with state law, since it doesn't appear to follow visible features.

Unless the SBOE has promulgated some sort of regulation that is contrary to statute, vesting authority in CBOE for defining election districts, this is pretty clear cut.

The simplest solution may be for the Common Council to pass a resolution defining the election district boundaries.  Except for the boundary between 5-1 and 5-2, they should conform to the ward boundaries.   If I were doing it, I would clarify which addresses are in:

Ward 1/Ward 2 along Front Street.
Ward 2/Ward 4 in the area where 3rd street stops.
Ward 4/Ward 5, including the split on Clinton, the incomplete block of 5th-Washington-Short-(Clinton), the east and south side of Harry Howard for its entire length; the houses on Lucille and Mill at the turn of Harry Howard, and the Fireman's home.

[jimrtex]

Are those the only two election board members? What official role do they have in NY?

Board of Elections in New York have two commissioners, one chosen by each major party (top two gubernatorial parties).  Larger counties may have four commissioners, and NYC has its own structure.  The statutes describe it as a board, even though it may only have two members.  The board hires clerks, voting machine technicians, custodians, etc, and election inspectors, and poll clerks.  They are in charge of registration and conducting elections, including local elections.  But it appears that in the that capacity they are acting on behalf of the cities and towns.

[jimrtex]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

The Board of Elections has two members, designated as commissioners.  The Board of Elections is in charge of registration as well conducting elections.

It appears to be common practice to have explicitly partisan staff.  Columbia County's BOE has two Deputy Commissioners, two elections specialists, and two voting machine/HAVA specialists, one of each party.  My guess is that "deputy commissioner" is actually a staff position.

[jimrtex]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

Yes, they have staff. Other than the Columbia triangle, most of the errors occurred with respect to the 2010 census. At about that time, the city clerk used the erroneous BOE map rather than the city charter to count ward populations (and did some block allocation errors), which map the BOE said was just an approximate guide and not exact (somebody is hiding something),  and Crosswinds was built, nobody is explaining why the voters were put in the wrong ward, other than three houses on that side of the street were in the wrong ward, and apparently had been for years for unknown reasons, and body bothered to check either the city charter or the erroneous map.  So either the BOE are sort of incompetent drones, or somebody was playing ball with what certain politicians wanted (and still want). We will never know probably.

I suspect that part of the error results from the definition of VTDs, which began with the 1980 Census.  PL 94-171 (passed by the 94th Congress in 1975) actually provides for the Census Bureau to tabulate data on a basis that is useful for states when apportioning legislative bodies, more so than what is now thought of the purpose of PL 94-171 (early releases of racial data to the census block level).  In the 1970s Alabama reapportionment litigation, one of the issues was that the court did not respect election precincts, since they only had enumeration districts, and didn't want to make estimates.  Instead, they told Alabama that they had to modify election precincts, and would love it.

VTD's, which have been known as election districts, voting tabulation districts, and voting districts, were introduced with the 1980 census.  They retain the acronym of VTD.

The Census Bureau doesn't define what they are, and not all states define them.  In 2010, two states did not, and two others only did so only in some counties.  In Maine, they somewhat match up with legislative districts when they divide cities or towns.  In Alabama, they vary quite a bit in what they represent.  An in issue in the current litigation is the number of VTDs that were split, but there also statements that a particular VTD had to be divided (because it was larger than a House district).  "Anniston VTD" has 40,000 persons, even though the election web site shows maybe 15-20 polling places in the city.  In other cases, it appears based on names, that several polling places were conflated into a single VTD (the name appears to be the polling places separated by slashes (/).

The definition of VTD is a cooperative program between the Census Bureau and the States.  The Census Bureau sets the criteria and then determines if a state program is consistent with that.  So it was likely that there was some communication between the SBOE and the Census Bureau, and then between the SBOE and CBOE's.   The CBOE's might not have paid much attention, or simply sent their best paper maps.  Somewhere along the line, they would have noticed that the ED boundaries didn't follow visible features and merged EDs.  Who knows where these decisions were made (Census Bureau, SBOE, or CBOE?).  It was likely not an executive decision.  The recognition by the Census Bureau was probably made by a computer.

The VTDs for 1990, 2000, and 2010 in Hudson are the same.  I don't know about 1980, or even if New York participated in 1980.

Hudson used to have seven election districts, with wards 3 and 5 divided into two ED's each.  The two ED's in ward 3 were merged fairly recently, I've seen election results online with results from the two EDs.  The names of the VTD's reflect the ED's that they were composed of (ie, "1-1 2-1 4-1", "3-1 3-2", and "5-1 5-2").

The VTD's correspond to visible streets, but there are two errors.  One is the notch back from 5th Street to Short Street, which is on Prospect, rather than Washington.  I assume that someone misinterpreted a map, and thought that they were taking into account that a a Harry Howard-Washington-5th-Clinton block does not actually exist, and moved the notch a block south.

The other error was the Columbia Triangle.  The Common Council resolution defining the ward populations following the 2000 Census had the Columbia triangle in Ward 5 (this matches the charter).  I had assumed that Ward 3 did conform with the VTD, but the VTD population did not match the post-2000 resolution, and the discrepancy exactly matched the census block total.   But even with that adjustment, the "Ward 5" VTD did not match the post-2000 resolution.  The post-2000 resolution thus did follow the ward boundaries in the charter.

The Columbia Triangle is shown as being in Ward 5 in tax records.  But Mayor Hallenback was the (county) supervisor for Ward 3, so the boundary may have been accidentally shifted some time before then.

The street one block north of Warren was Diamond Street.  It was Hudson's red light district (IIRC the 300 block).  The 1891 atlas shows it as Diamond Street, with the small diagonal segment going northeast being Columbia St to where Union Turnpike and Columbia Turnpike came together.  Since the alignment is closer to that of the Union Turnpike, they may have wanted to emphasize that the street was the connection to the Columbia Turnpike rather than part of the Union Turnpike.

Presumably when Diamond Street was renamed to Columbia Street, the segment of Union Turnpike was included in the renaming.  I don't know why the renaming occurred, but it might have something to do with creation of highway routes.  9G is along Columbia, rather than Warren.  Before there was signage, it might have been considered confusing to have the street name changed.   A route guide might have advised motorists to stay on 3rd Street when entering Hudson from the south and then make a right turn on Columbia Street.   US 9 originally took Prospect Ave to a short stretch of Columbia Street, before turning north on Fairview.  The current alignment is much more complicated.

In rural areas, roads are not good precinct boundaries, since they likely form the connection to the polling place.  It would be natural to designate both sides of a road as being in a single precinct, just as they would be for a school bus route, mail route, etc. 

The houses south of Harry Howard, just north of Underhill Pond were there since at least 1929 (they are shown on topo maps), and they are in fact in Ward 4.  If any of the houses on Lucille and Mill street at the turn of Harry Howard existed then, they would have been in Ward 4.  And the Firemen's Home was in Ward 4.  It would not be unnatural to believe that all houses on Harry Howard were in Ward 4.   The same sort of rule may be the norm in the other portions of the county.  Since town officials are elected at large, election districts are solely for the convenience of voters (and election officials).  If the BOE came to believe that all Harry Howard addresses were in Ward 4, it would be natural to add the couple of houses on the east side of Harry Howard further out, and then the new apartments.

Meanwhile the VTD delineations would come in handy if you had to publish a map on the web.   Until the mid-90s, there would be no need for maps other than a paper map to be sure voters were registered in the correct ED, and perhaps for poll workers to check on election day when a lost voter appears.  They could show them the map, and direct them to the Sleepy Hollow Town Hall.

[muon2]

Is the Torie plan still slated to go before the voters later this year? If it is and it is approved, does that force the BOE to make the requisite changes to the election districts?

[jimrtex]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

Yes, they have staff. Other than the Columbia triangle, most of the errors occurred with respect to the 2010 census. At about that time, the city clerk used the erroneous BOE map rather than the city charter to count ward populations (and did some block allocation errors), which map the BOE said was just an approximate guide and not exact (somebody is hiding something),  and Crosswinds was built, nobody is explaining why the voters were put in the wrong ward, other than three houses on that side of the street were in the wrong ward, and apparently had been for years for unknown reasons, and nobody bothered to check either the city charter or the erroneous map.  So either the BOE are sort of incompetent drones, or somebody was playing ball with what certain politicians wanted (and still want). We will never know probably.

The 1940 Census has the correct ward boundaries.  Union Turnpike had been renamed to Columbia Street by 1932, the date of the map prepared by the Department of Public Works, City Engineer J. McClure Harold (last name is fuzzy).  The map has a stamp from 1938, and the ward boundaries and enumeration districts were added with colored pencils for the 1940 Census.

The property tax records have the correct wards.

Benjamin Murell was a Ward 4 alderman from 1962-1974, and Ward 4 supervisor from 1974-2003 when he died.

There are two (or three) issues which may be getting conflated here.

(1) Voters voting in the wrong ward.
(1a) Voters on the east side of Harry Howard with Harry Howard addresses (2 or 3 houses) voting in Ward 4.
(1b) Voters in the Crosswinds Apartments voting in Ward 4.
(1c) Voters in the Columbia Triangle voting in Ward 3.

(2) Errors in calculating the base populations used in calculating voting weights.
(2a) Area between Prospect St. and Harry Howard betwen Short/Harry Howard and 5th St 5th Street Extended, which are in the "Ward 5" VTD, but are actually in Ward 4, and apparently vote in Ward 4.
(2b) Possible small errors in allocating population between Wards 2 and 4.
(2c) Misallocation of Front Street block.

(3) Relocation of the Firemen's Home.

Matt Murell could not have campaigned with his father at Westwind Apartments (1b) since it was built after his father died.  It is possible that he campaigned at the two or 3 houses on the east side of Harry Howard (1a).  If so, they were erroneously voting in the wrong precinct.  But there must have been only a handful of voters, and it plain weird to be in a different ward than all of your nearest neighbors.

Matt Murrell might have campaigned with his father along the houses along the Harry Howard just north of Underhill Pond, or in the area bounded by Harry Howard/Short-Prospect-5th St-Clinton(extended) to the south of Underhill Pond (2a).  But these areas are not subject to correction on the voter rolls.

Matt Murrell might have campaigned with his father at the Firemen's Home, but never in the current building, since it was also built after his father died.

Does the Board of Elections have any old registration rolls?

[jimrtex]

Can someone help>  Go to:

1940 Census Records

Click on Getting Started.  Below "Do you know the 1940 enumeration district number?" on the far right side, enter "New York" and "11-35"

You should see a thumbnail of a census form.  Click on it, and you should see image 1 of 28.  See if you can jump to Image 15 (in upper right corner).

=================================

Which ward are 496, 498, and 500 Clinton registered in on the voting  rolls?

[muon2]

One issue in all of this is the assignment of parcels and residences that are split by political lines. It happens frequently in IL. Consider this gem from the Chicago suburbs.



The orange line is a current boundary between state representative districts. The red line is a boundary between state representative, senate and congressional districts. The subdivision sits entirely within a single village, but spans a township line which was used as the red boundary line. Since election precincts are assigned by township, it might make some sense that a line follow the township line and not local streets.

Initially townships were divided into one square mile sections, and then the sections into quarter-sections a half mile on each side. Those section lines are used in survey references and in some parcel descriptions. When subdivisions are built the parcels are typically designated based on the recorded plat of subdivision and not based on the section, though surveys will reference back to the (quarter-)section. The orange line follows one of those old section lines which isn't a precinct line.

In most cases when a subdivision comes in, the census blocks are adjusted for the next decennial census to follow streets, rather than section lines. The exception is when a political line already follows that section line and then census blocks are defined splitting subdivision blocks along the section line. That tends to perpetuate into future censuses, and then those lines can get used for future political boundaries as in the map above.

The difficulty is in the assignment of residences to political districts given the splits that occur. In IL the residence is associated with its address which geographically is generally the main door of the building. Thus someone may actually sleep in one district, but officially reside in another based on the location of the address door. Needless to say it is easy for both the Census and local election officials to miscode these locations and I've caught quite a few such errors over the years. To complicate assignment, school district residence is based on the location of the master bedroom in a residence.

The Hudson ward issue may require knowledge of how individuals are assigned residence, and I don't know if NY uses the same type of assignment as IL.

[Torie]

It sounds like the map errors occurred at different times, and the boards at those separate times reacted differently to the errors.

Does the board have a staff, or do they actually maintain the voter lists themselves?

Yes, they have staff. Other than the Columbia triangle, most of the errors occurred with respect to the 2010 census. At about that time, the city clerk used the erroneous BOE map rather than the city charter to count ward populations (and did some block allocation errors), which map the BOE said was just an approximate guide and not exact (somebody is hiding something),  and Crosswinds was built, nobody is explaining why the voters were put in the wrong ward, other than three houses on that side of the street were in the wrong ward, and apparently had been for years for unknown reasons, and nobody bothered to check either the city charter or the erroneous map.  So either the BOE are sort of incompetent drones, or somebody was playing ball with what certain politicians wanted (and still want). We will never know probably.

The 1940 Census has the correct ward boundaries.  Union Turnpike had been renamed to Columbia Street by 1932, the date of the map prepared by the Department of Public Works, City Engineer J. McClure Harold (last name is fuzzy).  The map has a stamp from 1938, and the ward boundaries and enumeration districts were added with colored pencils for the 1940 Census.

The property tax records have the correct wards.

Benjamin Murell was a Ward 4 alderman from 1962-1974, and Ward 4 supervisor from 1974-2003 when he died.

There are two (or three) issues which may be getting conflated here.

(1) Voters voting in the wrong ward.
(1a) Voters on the east side of Harry Howard with Harry Howard addresses (2 or 3 houses) voting in Ward 4.
(1b) Voters in the Crosswinds Apartments voting in Ward 4.
(1c) Voters in the Columbia Triangle voting in Ward 3.

(2) Errors in calculating the base populations used in calculating voting weights.
(2a) Area between Prospect St. and Harry Howard betwen Short/Harry Howard and 5th St 5th Street Extended, which are in the "Ward 5" VTD, but are actually in Ward 4, and apparently vote in Ward 4.
(2b) Possible small errors in allocating population between Wards 2 and 4.
(2c) Misallocation of Front Street block.

(3) Relocation of the Firemen's Home.

Matt Murell could not have campaigned with his father at Westwind Apartments (1b) since it was built after his father died.  It is possible that he campaigned at the two or 3 houses on the east side of Harry Howard (1a).  If so, they were erroneously voting in the wrong precinct.  But there must have been only a handful of voters, and it plain weird to be in a different ward than all of your nearest neighbors.

Matt Murrell might have campaigned with his father along the houses along the Harry Howard just north of Underhill Pond, or in the area bounded by Harry Howard/Short-Prospect-5th St-Clinton(extended) to the south of Underhill Pond (2a).  But these areas are not subject to correction on the voter rolls.

Matt Murrell might have campaigned with his father at the Firemen's Home, but never in the current building, since it was also built after his father died.

Does the Board of Elections have any old registration rolls?

They are bootstrapping over those two or there houses, to claim that the Crosswinds is really in the 4th ward. I know this because of the key players behind the curtain, Rick Scalera, referenced that when making his case to me that their must have been some law passed not included in the City Charter defining the Ward lines, taking in the east side of Harry Howard Avenue. Needless to say, I was not impressed with "his case."

With reference to Muon2's question, if a house lot is split, the BOE can assign to either district. If it is a congregate living structure, or one apartment building that is spit, one is to assign the residents of the entire structure to the district where a majority of the residents live. I guess if where the majority of the residents live turns on where just a few residents live, because their apartment or room is split, than the BOE can assign all of the residents of the structure to either district. The Dem Commissioner explained all of this to me, and as to structures with multiple units, she has a written opinion from the State BOE I believe up in Albany.

[muon2]

It's interesting that voter assignment in NY is left up to the local officials for split residences. That's very different from my experience where the statute gives explicit direction to the local election officials, which makes it easier to bring mistakes to their attention for correction.

[Torie]

Hey, Jimtex, that 1891 map you found actually erroneously marks the Columbia triangle in the 3rd ward. Virginia Martin pointed that out. Can you post that 1938 map to which you allude?

[jimrtex]

Hey, Jimtex, that 1891 map you found actually erroneously marks the Columbia triangle in the 3rd ward. Virginia Martin pointed that out. Can you post that 1938 map to which you allude?

1940 Census Records

Click on Getting Started,
"Do you know the location where the person lived?"  Click on Start Your Search
On the left select the Location radio button, and then New York, Columbia, Hudson click on Search,
Select the one map.

Direct Link

You should be able to download this (as a jpg)

The underlying map was prepared by the city planning department and is dated 1932 (lower left of map).  The 1938 stamp is over on the right side.  My guess si that the city would have stamped it when they gave the copy to the Census Bureau.  Or perhaps it is a receipt stamp.  There might be an original, without the census markups in the city records.

By 1932, Diamond Street had been renamed Columbia Street, and the name had also been extended to the Union Turnpike portion.

Notice that the Firemen's Home and the reform school were separate enumeration districts, though noted as being in their correct ward.   When it was originally built, the prison was for girls.  In census records of that time, there is an extreme overage of females 15-18 or so.

I've been looking through the census sheets and have found some interesting things.

Enumeration District 11-35 is between 5th and 6th, north of State St, in the 5th ward, but then expands out.  On sheet 17 it includes 500, 498, and 496 Clinton (the enumeration was east to west down the north side of Clinton).   These are presumably the houses on the north end of 5th Street where it dead-ends at Clinton).   So it appears that they were counted in Ward 5.

Beginning on sheet 17, the enumeration proceeds up 6th street onto Glenwood.  Sheets 18 to 25 are for Glenwood, Parkwood, Oakwood, and Fairview, but nothing further north.  This indicates that the area east of Harry Howard was developed after WWII.

Enumeration District 11-33 is between 3rd and 5th street, north of State St, in the 4th Ward (but does not include the Firemen's Home).   On sheets 10 through 12, enumeration is on Harry Howard, proceeding through 84, so it includes the houses to the north of Underhill Pond.  But then there is one house with "(None)" for the address.

But then on Sheet 14, there are two more houses on Harry Howard with "(None)" for the address.  But there is also the notation "5th Ward".  These might be the houses on the east side of Harry Howard.   One of the occupants is Delaney, so possibly related to the current alderman?

[jimrtex]

One issue in all of this is the assignment of parcels and residences that are split by political lines. It happens frequently in IL. Consider this gem from the Chicago suburbs.



The orange line is a current boundary between state representative districts. The red line is a boundary between state representative, senate and congressional districts. The subdivision sits entirely within a single village, but spans a township line which was used as the red boundary line. Since election precincts are assigned by township, it might make some sense that a line follow the township line and not local streets.

Initially townships were divided into one square mile sections, and then the sections into quarter-sections a half mile on each side. Those section lines are used in survey references and in some parcel descriptions. When subdivisions are built the parcels are typically designated based on the recorded plat of subdivision and not based on the section, though surveys will reference back to the (quarter-)section. The orange line follows one of those old section lines which isn't a precinct line.

In most cases when a subdivision comes in, the census blocks are adjusted for the next decennial census to follow streets, rather than section lines. The exception is when a political line already follows that section line and then census blocks are defined splitting subdivision blocks along the section line. That tends to perpetuate into future censuses, and then those lines can get used for future political boundaries as in the map above.

The difficulty is in the assignment of residences to political districts given the splits that occur. In IL the residence is associated with its address which geographically is generally the main door of the building. Thus someone may actually sleep in one district, but officially reside in another based on the location of the address door. Needless to say it is easy for both the Census and local election officials to miscode these locations and I've caught quite a few such errors over the years. To complicate assignment, school district residence is based on the location of the master bedroom in a residence.

The Hudson ward issue may require knowledge of how individuals are assigned residence, and I don't know if NY uses the same type of assignment as IL.

PL 94-171 essentially requires the Census Bureau to tabulate on the basis of political subdivisions.  The text of PL 94-171 is pretty interesting to read.  I had expected something else.

In the 1990 Census, there were "census blocks" and "census tabulation blocks", which were subdivisions of census blocks.  The census blocks were bounded by visible features, while the census tabulation blocks could be tabulated for populations of political subdivisions.  The 1990 census would have split visible blocks on county lines, since census blocks are nested within census tracts, which are nested within counties.

For 2000 there were census blocks, which could have both visible boundaries and political subdivision boundaries.

For 2010 the Census Bureau did permit non-visible boundaries for VTDs, grudingly. 

The program for (re)defining VTDs appears to be based on coordination between the Census Bureau and a state liaison (likely appointed by the governor).  Presumably this would either be a statewide planning agency or demographic agency.   The challenge is that those with local knowledge and concern for detail don't work for the state.   How good the implementation is would depend on how well the state agency coordinates with county planning authorities.   And with VTD's you would also need lateral coordination with the SBOE and CBOE's.

I see in an earlier note that New York did not participate in the VTD program for 1980.  So the VTD's would have been defined in the late 1980s.