Sarah Ligon (@ligonish)
November 2021
In September 2021, I joined a team of two other graduate researchers assessing NYC’s newly-federally-funded Housing Mobility Demonstration for the NYU School of Law and School of Public Policy’s joint seminar on “Land Use, Housing, & Community Development in New York City”. The code, analysis, and visualizations in this repository are my own original contributions to the data section of what grew to a 56-page collaborative paper.
I was skeptical of housing mobility initiatives’ ability to deliver racially equitable housing choice to contemporary New Yorkers, and only grew more so while exploring the data below. The “root shock” and loss of community networks often experienced by Black and brown mobility program participants is well-documented, from early voucher recipients such as Chicago’s Gautreaux to the trauma of mass displacement over decades of New York officials’ policies of urban “renewal”. NYC landlords rampantly discriminate against voucherholders of color: in the spring of 2021 — just as NYC was receiving its millions of dollars in federal funding for more Section 8 Housing Choice Vouchers as part of the demonstration my team would evaluate — New York’s Fair Housing Initiative filed a lawsuit against 88 NYC landlords and brokers, alleging sweeping patterns of illegal voucher discrimination.
The analysis below, then, asks the 2017 NYCHVS to tell us about mobility patterns not just of New Yorkers in general, but those of color in particular. I also concentrated on what the NYCHVS could reveal about New Yorkers who already receive mobility-focused City housing assistance in the form of Section 8 HCV users, since that’s what the newly-funded Demonstration would expand. As indicated in the charts and figures below, patterns in the NYCHVS indicate Black, Latinx, and HCV-holder households experience very different access to housing mobility than the rest of the city.
My thanks to NYU Law teammates Eliza Ezrapour and Justin Cook for their brilliance and great good humor during what was, by any standards, an eerie semester.
Residential mobility is infamously difficult to quantify: publicly-available datasets allow very limited tracking of voluntary and (especially) involuntary household moves, with significant sampling challenges, standard errors, and tracking lags baked in.
For this project, my primary source of large-scale NYC housing mobility data was the U.S. Census Bureau’s “New York City Housing and Vacancy Survey: 2017 Data Files”, accessed 20 October 2021 and analyzed in Stata.
- See data_raw repository folder for original microdata records of household-level vacant, household-level-occupied, and person-level NYCHVS 2017 survey estimates.
- See documentation repository folder for NYCHVS codebooks, data dictionaries (including Sub-Borough Maps & Codes), and official Guide to Estimating Variances.
Pre-analysis setup workflow as documented in this repository’s Stata .do files:
- Download the three original occupied & vacant NYCHVS microdata files
- Make a fourth by appending vacant to occupied housing records (for housing unit estimates)
- Merge PUMA names for future spatial analysis with IPUMS microdata
- Convert final and replicate weights (in raw form they have five implied decimal places, so divide by 100k)
replace fw = fw/100000
foreach x of var fw1 - fw80 {
replace `x' = `x' / 100000
} - Set survey design parameters for accurate weighted estimates. I used Stata’s -svset- function, but you can also use R’s survey package or similar. See this UCLA tutorial on working with sampling weights in Stata.
svyset[pweight = fw], vce(sdr) sdrweight(fw1-fw80) fay(.5)mseThe stata_do_files repository folder holds replicable code to produce the Stata graphs and tabular data .csv exports I then formatted in Google Sheets and/or Datawrapper below. (Current Me would use R for the whole process, but 2021 Me was making myself learn Stata for a separate Wagner project.)
At first glance, New York owners and renters may seem unexpectedly static: while high-income Manhattanites’ headline-grabbing COVID-era relocations become something of a national curiosity in the year leading up to this project, NYC residents tend to live in the same units for unusually long periods of time. The most recently-available nationwide American Housing Survey (AHS), taken in 2017, places U.S. homeowners in their current residences for an average of 16 years, and renters for 4.5. In the same year, the 2017 NYCHVS suggested the average NYC owner has lived in the same house for 19.3 years, and the mean New York City renter household has lived in their current unit for 11.8 years — almost three times the national average for other American renters.
As indicated in Figure 1, there is little variance from this 11 to 12 year renter mean at the between-borough level, with the exception of Staten Island’s 7.9 year average rental duration (though, at 65.5% owner-occupied units, Staten Island is also the only borough where renters constitute the minority). At the between-household and between-district level, however, crucial differences in New York City renters’ degrees of mobility emerge.
Renters using Section 8 mobility vouchers are significantly more static than non-Section 8 renters across the city. Averaging 13.8 years, Section 8 voucher holders’ rental tenure is also close to that of Black New Yorkers (13 years), and almost as long as the 14-year average tenure of renter households below 100% of the federal poverty limit for their household size. White renters, on the other hand, spend a shorter-than-average number of years in their units, suggesting they move more often.
The NYCHVS asks heads of households when their household moved to their current unit, why they moved to their unit, and the name of the borough where they previously lived.
The most commonly-stated explanation among renters across the city was change in employment status (15%), followed by need for a larger dwelling unit (13%). In each borough, 5 to 6 percent of renter households said they moved to their current location in search of greater housing affordability, while 1 to 3 percent reported moving due to eviction, displacement, or harassment by a landlord.
There’s insufficient data at the sub-borough level to produce accurate weighted estimates of post-eviction/ displacement/harassment moves to all 55 New York City sub-boroughs. Sampled NYCHVS survey respondents who reported moving for this reason were, however, sufficiently clustered within just two neighborhoods to indicate — at the 95% level of statistical significance — that 8.8% of self-reported evicted/harassed tenants moved to units in the Morrisania/East Tremont sub-borough of the Bronx (which approximately encompasses Community Districts 203 and 206), and 4.4% to Washington Heights/Inwood (Community District 112).
Households also tended to cite a need for greater affordability fairly evenly across boroughs — an unsurprising fact given both the general phrasing of the survey question and the broad spread of households below federal poverty limits in virtually every neighborhood of New York City.
A group that does, however, appear to experience moving patterns more restricted than the rest of New York City’s is the one meant to have the broadest range of housing choice: Section 8 Housing Choice Voucher (HCV) holders, over 27% of whom are clustered in just four sub-boroughs in the Bronx.
These four sub-boroughs, where 27.29% of all NYC’s Section 8 HCV recipients could use their vouchers, are predominantly Black and brown. Resident HCV users reported high levels of heating breakdown and other maintenance failure by landlords.
NYCHVS’s estimated 145,499 Section 8 voucher holder households appear unusually static when taken as a whole (see Figure 1 above). But HCV-holder households of different races experience very different levels of housing stability.
Black householders — who overall average the lengthiest unit stays citywide (15.3 years) when compared to other New York householders of all other races — experience the shortest duration in Section 8 units in relation to all other HCV users (see Fig. 3).
These discrepancies also extend to Section 8 voucher holders’ moving patterns across boroughs and sub-boroughs: 43% of Black voucherholders move to the Bronx, while 72% of white voucher recipients move to Brooklyn (see Table 4).
Among Section 8 voucher households across NYC, 81.2% of white voucher holding households move to a unit within their current borough, while only 67.6% of Black Section 8 households and 69.1% of Hispanic/Latinx households use their vouchers within the borough in which they already reside.
For a broader view across the city, here’s a summary table of some mobility-related housing factors broken down by Community District boundaries.
