The upcoming primary on Thursday, September 13, 2012 for New York State Senate and Assembly will be the first election using state legislative districts drawn during the 2012 redistricting process.
The Center for Urban Research at the CUNY Graduate Center has updated its interactive maps of these districts (see screen shot at right), highlighting the differences in race/ethnicity characteristics between total population and voter-eligible population - in other words, comparing the characteristics of all those who live in the new districts versus the smaller group who will be eligible to vote for each district's representatives. In some cases the differences are striking.
An overview plus district-by-district data is presented below.
Note that the discussion below refers to "race/ethnicity groups" based on the major race and Hispanic categories used by the US Census. We use the Census Bureau's mutually exclusive race and Hispanic origin categories as follows: 1) non-Hispanic White; 2) non-Hispanic Black; 3) Hispanic/Latino; 4) non-Hispanic Asian; & 5) all other persons (representing non-Hispanic people who checked two or more races, Native Americans, Hawaiians, or some other race). Therefore, when we refer to White, Black, Asian, and others, we are referring to non-Hispanics. People who identified themselves as Hispanic or Latino on the census form can be of any race.
Everyone is represented, but not everyone can vote
Legislative redistricting rests on the principle of one person, one vote; that is, each legislator should represent nearly the same number of residents. While each legislator represents all the people in his/her district regardless of whether or not they can vote, however, they may be more responsive to those who are eligible to vote.
Recent Census Bureau estimates of eligible voters (citizens age 18 and older) reveal that the share of the resident population who can vote is not distributed evenly across recently redistricted Senate and Assembly districts. (See data sources below.)
In both houses of the state legislature, upstate districts generally have substantially higher shares of eligible voters than New York City districts. Almost three-quarters of the upstate population is eligible to vote, but the share is only three-fifths in New York City. And the share is even lower in 12 of the state's 63 Senate districts, all in New York City. Of the 150 Assembly districts, 37 have an eligible voting population less than 60% and all but five are in New York City.
Eligible voters make up the lowest share of the population in Senate District 13 in Queens, with only 43%. Assembly District 39 overlaps the lower portion of Senate district 13 and has an even lower share of eligible voters (the lowest in the state): only 34% of its population can vote.
The charts below show the components of district population:
- dark blue bars correspond to non-citizen adults (not eligible to vote),
- red bars
represent the population under age 18 (not eligible to vote),
- green bars show eligible voters (citizens of voting age).
The bars are grouped to show the districts on Long Island on the left, New York City in the middle, and upstate New York on the right. The bars are sorted lowest to highest within each region by each district's number of non-citizen adults.
For each chart, hover your mouse over the bars to display the district number, population of each segment, and total population of the district. You can see each of these districts along with the racial/ethnic breakdown of eligible voters vs. total population at CUR's interactive map.
Figure 1: Senate district population sorted within each region by number of non-citizen adults
Figure 2 below presents the same information but as percentages -- the height of each bar represents 100% of the district's population, and the size of each blue, red, or green segment of the bar corresponds to the district's share of each category.
Figure 2: Percent of each Senate district's population by eligible/not eligible to vote category
Figures 1 and 2 show that with few exceptions, upstate districts have the highest shares of eligible voters because they have the smallest shares of non-citizens (the size of the youth population is generally consistent across most districts). In statistical terms, the Pearson correlation coefficient between non-citizen adults and the share of eligible voters is -0.92, indicating a very strong negative relationship (the greater the number of non-citizen adults, the lower the share of eligible voters). The correlation coefficient between the population under 18 and eligible voters is not as strong -- only -0.55.
In two city districts the share of eligible voters is especially low. Specifically, in Senate district 13 and district 33 (part of the Bronx), less than half of the population is eligible to vote. In district 13 this is primarily due to non-citizen adults, but in district 33 it is due more to a large population under age 18. But it is possible that a substantial portion of the youth population in this district -- where Latinos make up almost 70% of the district -- are also not citizens.
There are some exceptions in New York City. On Staten Island, just under 74% of the population in Senate District 24 is eligible to vote (covering the southern portion of the borough, with very few non-citizens). In Manhattan, almost 75% of Senate district 28 (Upper East Side and Murray Hill) and 80% of Senate district 27 (Chelsea, Greenwich Village, the East Village) are citizens of voting age. These districts each have a substantial share of non-citizen adults, but the youth population is relatively small.
Assembly districts provide a more mixed picture, with some districts much worse off in terms of eligible voters and some districts better off. See Figures 3 and 4 below.
Figure 3: Assembly district population sorted within each region by number of non-citizen adults
Figure 4: Percent of each Assembly district's population by eligible/not eligible to vote category
Implications for minority representation
These disparities may have important electoral implications: voters in districts with a small share of eligible voters have a greater say in who is elected. Instead of being 1 out of 300,000 people who can vote, they are 1 out of 150,000, for example. This may not be a great concern where the interests of eligible voters are aligned with those who cannot vote, but the characteristics of the two populations may be very different in districts with large populations of non-citizen adults.
For example, several districts in New York City and its suburbs were redistricted this year to encompass overall majority or plurality Asian or Hispanic populations, but their eligible voters remain substantially or predominantly White. If candidates in these districts focus their campaigns on a set of issues of interest to potential voters but less important to a substantial bloc of non-voters, this may heighten the sense of disenfranchisement among district residents who are not yet able to vote. This may also provide them with an incentive to naturalize and register so their voices can be heard more fully at the ballot box.
For example, the redistricted State Senate district 16 has been described in the media as an "Asian-majority" district: its population is 53% Asian American and less than a quarter White (and the combined Black/Hispanic share is 20%). It therefore might be considered a likely seat for an Asian American candidate to win.
However, based on Census Bureau estimates, Asians make up only 40% of that district's eligible voters while the White share of eligible voters increases to just over 36% and the combined Black/Hispanic share is 21%. The racial/ethnic makeup of the eligible voters indicates this would more likely be an "Asian opportunity" district rather than a "safe" Asian district. In the upcoming Democratic primary in district 16 there are only two candidates, both White, one of whom has the backing of prominent Asian politicians. The sole candidate in the Republican primary is Asian American.
Of course many factors shape election outcomes, including the field of candidates, the issues that are important to local voters, and how closely people vote along race/ethnicity lines. Further, the number of registered voters will differ from (and likely be less than) the number of eligible voters, and the number of people who actually vote will be even smaller. Unfortunately, the voter registration process does not record the race and ethnicity of registered voters, so the American Community Survey estimates of the race/ethnicity composition of eligible voters are the most comprehensive data available to determine the differences between the eligible electorate and total resident populations in the new State Senate and Assembly districts.
Examples of districts with these characteristics in the State Senate include (links open in a new window):
- District 12 (Queens) – This district has a slightly larger Hispanic population in number and percent than the White population (there are 116,752 Hispanics, 37% of the district, compared with 114,601 Whites, 36%), but Whites represent almost half the district's eligible voters (49%, or 86,034 people) and Hispanic eligible voters drop to 30% (only 53,204 people).
- District 16 (Queens) – This district has an Asian population majority (53%), with 25% of residents being White and 20% Black or Hispanic, but the Asian share of eligible voters drops to 40% while the White share increases to 36% (and the Black/Hispanic share rises modestly to 21%).
- District 29 (the incumbent in the old district 29 is not running, so technically this is an open seat, though it was substantially redrawn to include only small areas in Manhattan but is mainly in the Bronx) - Hispanics make up 53% of the district's total population (almost 168,000 people), but this drops to 41% when limited to eligible voters (just under 74,000 eligible Hispanic voters). In contrast, Whites are only 20% of the total population (64,059 people), but make up 30% of eligible voters (just under 53,000).
- District 31 (Manhattan) - Hispanics represent a majority of the district population (56%, almost 180,000 people) and Whites are only 29% (92,000 people). But the Hispanic share of eligible voters drops to 45% and the White share increases to 39%, a much closer gap.
- District 35 (Westchester County) - Whites represent 43% of this district's population and Hispanics are 31%. But Whites are 56% of eligible voters and Hispanics only 18%.
Examples in the Assembly include:
- District 6 (Suffolk County) - Hispanics represent almost 60% of the population while Whites are only 20%. Hispanics still have a greater share of eligible voters (41%) compared with Whites (33%), but the difference in number of eligible voters for both groups is small: 26,400 eligible Hispanics compared with almost 22,000 Whites.
- District 18 (Nassau County) - close to half this district's population is Black (47%) and Hispanics represent 38%. But the Black share of eligible voters increases to 58% (almost 41,000 people) and the Hispanic share drops to 20% (13,000 people).
- District 25 (Queens) - an open seat, the district's population is majority Asian (52%) and only 29% White. Hispanics and Blacks combined represent 16% of the population. But the shares of eligible voters for Asians and Whites are equal: both make up 40% of eligible voters, just over 30,000 people each. The combined Hispanic/Black share of eligible voters increases slightly to 17.5%.
- District 40 (Queens) - a substantial Asian majority (62% - the district mainly covers the Flushing neighborhood) with Whites only making up 17%. But the Asian share of eligible voters drops to 46% while the White share increases to 31%.
We use two data sets for comparisons in our maps and analysis,. For total district population, we use 2010 decennial Census data as modified by the New York State Legislative Task Force on Demographic Research and Reapportionment (LATFOR) to account for the last known residence of prisoners in state facilities as of April 1, 2010. See LATFOR's summary of the data and the actual data sets and detailed documentation. This adjusted data was the basis by which Senate and Assembly districts were redrawn in 2012, per state law.
For the population of eligible voters, we use a special tabulation from the Census Bureau from the American Community Survey (ACS) for the 5-year period 2006-2010. This tabulation is referred to as estimates of the "citizen voting age population" (or CVAP). These are sample-based estimates averaged over a 5 year period, not a 100% count as taken by the 2010 Census. But the ACS provides the most recent data on citizenship and when combined with voting age it provides the only way of estimating those who are eligible to vote.
The Census Bureau provides CVAP estimates for areas no smaller than Census block groups; these are spatial aggregations of Census blocks (which are the smallest Census geographic units) from which legislative boundaries are constructed).
The block group data as provided by the Census Bureau does not include any explicit relationship to newly-redistricted State Senate or Assembly districts (i.e., there are no legislative district IDs corresponding to each block group). Therefore, in order to tally the CVAP estimates by legislative district, we used ArcGIS to spatially assign legislative district IDs to each block group based on the Senate/Assembly district that contained the block group. If a block group spanned multiple districts, it was only assigned to one Senate or Assembly district, depending on where its geographic center (i.e., centroid) was located. This introduces some degree of aggregation error into the results -- sometimes a block group centroid falls within one legislative district even though most of the land area or population of the block group is within the adjacent legislative district. Overall, any over- or under-estimation will average out across the state, and the misallocation of block group data should be minimal. We used a similar approach during the redistricting process, discussed in detail here, to allocate voter tabulation data from LATFOR to legislative districts.
When you visit CUR's interactive map, move your mouse over each district to display total population counts by race/ethnicity along with CVAP estimates for each race/ethnic group.
Support for CUR's demographic analysis and educational efforts especially related to redistricting this year primarily comes from the Hagedorn Foundation.
For the mapping application at www.urbanresearchmaps.org/nystatelegislature/map.html, David Burgoon, CUR's application architect, constructed and designed the site, with data analysis support and overall conception from CUR's Mapping Service director Steven Romalewski along with Joseph Pereira, director of CUR's CUNY Data Service.
The online mapping application application relies on geographic data hosting by cartoDB, open source mapping frameworks and services including Leaflet.js and Cloudmade, and ESRI's ArcGIS software for cartography and data analysis, as well as jQuery and a modified version of the jQuery Sparklines plugin for the district charts.