Assessing ATi a Indicators and Race/Ethnicity Paul M. Ong and Jonathan D. oa January 18, 2021 Ps a UCLA CENTER FOR Center for Neighborhood HEALTH POLICY RESEARCH. UCLA Knowledge = Acknowledgements The author is appreciative of the valuable input provided by Professor Vickie Mays and partial financial support from her center (UCLA BRITE) in the initial construction of the vulnerability index based on pre-existing health conditions, the UCLA Center for Health Policy Research for providing the tract-level CHIS (California Health Interview Survey) data, Professor Ninez Ponce and Jacob Rosalez for reviewing the draft, and Megan Potter for designing the brief. We also want to thank Vananh D. Tran, Richard Calvin Chang, Jacob Rosalez, and Ninez A. Ponce for assisting in preparing the analysis for presentation. The Center for Neighborhood Knowledge at UCLA acknowledges the Gabrielino/Tongva peoples as the traditional land caretakers of Tovaangar (Los Angeles basin, So. Channel Islands) and pay our respects to the honuukvetam (ancestors), 'ahiihirom (elders), and 'eyoohiinkem (relatives/ relations) past, present, and emerging. Disclaimer The views expressed herein are those of the author and not necessarily those of the University of California, Los Angeles. The authors alone are responsible for the content of this report. Information about the Author Professor Paul M. Ong is the Director of the UCLA Center for Neighborhood Knowledge, and has a Ph.D. in Economics from the University of California, Berkeley. Jonathan D. Ong is a researcher at Ong&Associates and a graduate student in computer science at Arizona State University. Suggested Citation Paul M. Ong and Jonathan D. Ong. (2020) "Assessing Vulnerability Indicators and Race/Ethnicity." UCLA Center for Neighborhood Knowledge. Gerson Repreza Abstract This study assesses four vulnerability indicators that are being considered by public agencies as policy tools to select the most-at-risk neighborhoods for interventions. These indicators can play a role in prioritizing the provision of pandemic resources and services; consequently, they have implications for how many people of color and minority neighborhoods are served. The study compares three vulnerability indicators developed prior to COVID-19 and one developed in response to the pandemic. Two sets of assessments are conducted. The first calculates the degree of concordance between pairs of indicators, that is, how frequently they identify the same tracts as being disadvantaged. The analysis finds that low rates of commonality (approximately less than half of all designated tracts); therefore, the choice of indicator inherently translates into a significant variation in the tracts classified as being eligible or ineligible for prioritization. The second set of assessments examines the differences among the indicators by comparing the racial composition of the residents in designated high-vulnerability tracts, and by comparing the relative number of minority neighborhoods included in high-vulnerability tracts. The analyses find substantial differences among the indicators in population compositions and proportion of minority neighborhoods included. The findings can help ameliorate a policy dilemma. Despite the reality that African Americans and Hispanics have suffered disproportionately from COVID-19, the 1996 Proposition 209 prohibits the state from explicitly using race as a factor in the provision and distribution of pandemic relief and coronavirus vaccines. The study's findings provide insights into which of the four vulnerability indicators can serve as a reasonable proxy, one that captures an important underlying mechanism producing systemic racial inequality. By several criteria, among the indicators that do not explicitly include race/ethnicity as an input, the indicator based on pre- existing health conditions (medical vulnerabilities) performs best in including African Americans. A final recommendation is that public agencies should develop and construct new pandemic- oriented indicators to help guide policies beyond racial equity. Introduction This study assesses four vulnerability indicators that are being used as policy tools to select the most-at-risk neighborhoods for prioritized interventions. Without timely and geographically precise data on COVID-19 infections and the factors that determine the rate of infection, public agencies have turned to pre-pandemic vulnerability indicators that measure socioeconomic and other types of vulnerabilities. The underlying assumption is that these indicators are highly correlated with disparities in COVID-19 outcomes, thus are useful proxies to identify at-risk neighborhoods. The indicators can play a role in prioritizing the provision of pandemic resources and services. For example, both National Academies of Science, Engineering, and Medicine (NASEM), and the Center for Disease Control and Prevention (CDC) have recommended the use of two pre-existing indicators to identify the most-at-risk places to allocate a proportion of the vaccines,12 a practice that has been adopted by many states.3 This study compares three pre-pandemic indicators and a more recently developed indicator based on pre-existing health conditions. The analysis focuses on the numbers of people of color residing in designated high-vulnerability neighborhoods, and the relative number of minority neighborhoods that fall into the high-vulnerability areas. Race/ethnicity is an important factor because Latinos, and Pacific Islanders are disproportionately more likely to be infected by COVID-19, and African Americans, Latinos and Pacific Islanders are more likely to die from an infection.4 While Asian Americans have lower rates, there is evidence that some Asian ethnic subgroups also bear a disproportionate share of negative COVID-19 outcomes.5 Race is also important because people of color encounter multiple dimensions of inequality that are only partially reflected in the indicators. For example, none of the indicators include an input variable that directly measures systematic differences and disparities in policing. The assessment utilizes a policy-oriented exercise that simulates what would be the outcomes if an indicator is used to identify the most vulnerable neighborhoods. By comparing the hypothetical results for the four indicators, the policy-based exercise provides insights into which neighborhoods and populations would be classified as at high-risk and consequently prioritized for programmatic action. The findings show noticeable differences in the groups and places designated as being vulnerable, thus the choice of which indicator to use has highly consequential implications in terms of who is served and who is not along racial lines. Indicators and Method This study uses three pre-pandemic indicators at the census tract level. The social vulnerability index (SVI) was created by the CDC to identify vulnerable areas for disaster planning and response.® The indicator uses 15 ACS variables organized into four dimensions: socioeconomic status, household composition, minority status and language, and housing type and transportation. This study uses the most recent CDC version of the SVI, 2018.7 The second national indicator is the Area Deprivation Index (ADI), which initially developed by the Health Resources & Services Administration, and subsequently refined by Dr. Amy Kind and her research team.8 ADI uses ACS data related to income, education, employment, and housing quality, and designed to inform health delivery and policy. ADI block-group level data are aggregated to tracts using population weights. The SVI and ADI are the two indicators recommended by CDC and NASEM to identify vulnerable places. The Healthy Places Index (HPI) was developed by the Public Health Alliance of Southern California? and designed to help policy makers target the most disadvantaged communities for interventions and resources. California has adopted the HPI as one of its policy indicators. The index uses 25 variables from eight sources (including ACS) and organized in eight domains: economy, education, healthcare access, housing, neighborhoods, clean environment, transportation, and social environment. The rankings are inverted so higher values denote greater vulnerability. Neither the HPI nor the ADI include race/ethnicity variables. The study includes a newer indicator constructed specifically for the pandemic. The UCLA Pre- Existing Health Vulnerability (PHV) index captures the risks or severity of COVID-19 infection due to preexisting health conditions!° and is based on data from the California Health Interview Survey (CHIS)21. The input data for the PHV are small area estimates at the tract level modeled from CHIS data. PHV is a composite of six variables: diabetes, obesity, heart disease, overall health status, mental health and food insecurity. Each variable is ranked. and the variables are then summed at the tract level. The majority of the tracts are estimated directly from CHIS data, the values for other tracts are estimated with predictive regression models using data from ACS and CalEnvironScreen!2. The regression model also incorporates information on PHV from nearby tracts is also used when available. For the state as a whole, the indicators (ranked as percentiles) are highly correlated with each other, with unweighted r-values ranging from 0.69 (ADI and SVI) to 0.88 (HPI and SVI), and the population weighted correlations range is very similar from 0.70 (ADI and SVI) to 0.88 (HPI and SVI). These results are not surprising since all four indicators include ACS data, in some cases the same input variables. The correlations based on percentile rankings, however, do not provide sufficient information about the indicators' performance as a policy instrument that defines which neighborhoods are eligible and ineligible for assistance. The NASEM and CDC recommends, for example, that the most-at-risk places are those ranked in the top 25% in vulnerability. California also follows this practice, so for the simulation analyses, tracts are categorized as 1 if they are in the top quartile of the vulnerability ranking, else 0. We use two ranking procedures. The first is for all tracts in the state, which simulates how state agencies and statewide organizations might use the data. This can produce a disproportionately larger number of tracts in some regions relative to other regions. The second is separate ranking within each county (that is the top quartile within a given county), a procedure that county health departments might use. Two sets of assessments are conducted. The first calculates the degree of concordance between pairs of indicators, that is, how frequently they identify the same set of tracts as being disadvantaged relative to tracts that are uniquely by just one indicator. The concordance rate is the number of tracts in common divided by tracts identified by either or both indicators. Figure 1 provides an example using the SVI (on the Y or vertical axis) and the PHV (on X or horizontal axis). The black lines within the graph represent the 75th percentile for each indicator. The tracts identified by SVI as being highly vulnerable are in areas "A" and "C", and the tracts identified by PHV as being highly vulnerable are in areas "C" and "B". The concordance rate is the tracts in "C" divided by the tracts in "A" plus "B" plus "C". If there is extensive commonality in the way places are classified, then the policy choice of indicator is not critical. However, if there are significant differences, then the choice may matter. Empirically the former is the outcome; consequently, the indicators can potentially produce different results in terms of the populations and neighborhoods designated to receive priority because of high risk. Figure 1: Distribution of Tracts by SVI and PHV Statewide Ranking SVI and PHV by Statewide Ranking 100 90 80 70 60 40 a ae 30 og 20 10 0 10 20 30 40 50 60 70 80 30 100 The second set of assessments examines the differences among the indicators by comparing the demographic outcomes. The first step is based on population composition by race. This is done by summing up the counts from the 2015-19 American Community Survey for the tracts designated as being highly vulnerable. For the tracts included in the study (which mostly excludes places with no population or with a disproportionately large number of persons in group quarters), non-Hispanic Whites comprise 37.2% of the total, Asian Americans comprise 14.5% of the total, African American comprise 5.8%, and Hispanics 39.0%. Given that African Americans and Hispanics are relatively more disadvantaged, then it is likely that they comprise a larger percent of the population in high vulnerability tracts. The second step is calculating the probability that majority minority neighborhoods are classified as being vulnerable, using 2015-19 ACS data. Minority neighborhoods have disadvantages beyond those encountered at the individual level.13 Residents in these places face place-based forms of discrimination and unfair treatment by people, firms and institutions, and systematic barriers to regional opportunities. Operationalizing the designation is done by first identifying the tracts that are predominantly of one demographic group (African Americans, Asian Americans, Hispanic and non-Hispanic Whites).14 Among the 8,059 tracts included in the study, there are 2,937 tracts are majority non-Hispanic White, 373 tracts are majority Asian American, 66 are majority African American, and 2,522 Hispanic. The analysis determines the proportion of these majority tracts are designated as being highly vulnerable. Because there are a very large number of majority Hispanic tracts, the analysis also examines very poor majority Hispanic neighborhoods (defined as a majority Hispanic tract where persons below the federal poverty line comprise 30% or more of the total population). Results The two figures below summarize the concordance analysis. Figure 2 is based on designating the vulnerable neighborhoods by ranking for all tracts within the state. Because of regional differences in overall socioeconomic and other demographic characteristics, designated vulnerable tracts are not proportionately distributed among the regions. For example, 29% of all tracts are in Los Angeles County, while 12% of ADI vulnerable tracts, 40% of HPI vulnerable tracts, 32% of PHV vulnerable tracts and 40% of SVI vulnerable tracts are in the County. Concordance rates vary from a low of 37% (ADI and SVI) to a high of 61% (HPI and SVI). The average is 46%, meaning that there are more uniquely identified tracts than commonly identified ones; therefore, the choice of indicator inherently translates into a significant variation in the sets of tracts that are classified as being as being eligible or ineligible for prioritization. Figure 2: Concordance Rates of Tracts Based on Statewide Vulnerability Rankings Common Tracts Statewide Ranking 70% 60% 50% 40% 30% 20% 10% 0% ADland HPI ADlandPHV ADlandSVI HPlandPHV HPlandSVI SVIand PHV Figure 3 is based on designating the vulnerable neighborhoods by separate rankings within each county. This implicitly means that there are proportionally the same number of designated vulnerable tracts (the top quartile) in each county regardless of sizable differences in socioeconomic and other demographic characteristics. This can be seen in the case for affluent San Francisco County by comparing the designation by within county ranking and statewide ranking. By definition, approximately a quarter of San Francisco tracts are designated as being vulnerable (24% to 25%, depending on how tracts at the 25th percentile are classified). On the other hand, 1% or none of San Francisco tracts are designated as being vulnerable when using statewide ranking. Despite the differences when using within county rankings, the concordance rates are low, ranking from 43% (ADI and PHV) to a high of 59% (HPI and SVI), with an average of 49%. As with the previous analysis using statewide rankings, the results here also means that there are more uniquely identified tracts than commonly identified ones. The implication is the same: the choice of indicator inherently translates into a significant inconsistency in the tracts that are uniquely classified as eligible or ineligible for prioritization. Figure 3: Concordance Rates of Tracts Based on Within County Vulnerability Rankings Common Tracts, Within County Ranking 70% 60% 50% 40% 30% 20% 10% 0% ADland HPI ADlandPHV ADlandSVI HPlandPHV HPlandSVI SVIand PHV The following two tables summarize the population analysis by race. Table 1 is based on designating the vulnerable neighborhoods by ranking for all tracts within the state. As expected, African Americans and Hispanics comprise a larger percent of the population in high vulnerability tracts because these two groups are relatively more disadvantaged. (For comparison, the following are the percentages for all tracts in the study: 27% for non-Hispanic Whites, 14% for Asian Americans, 6% for African American and 39% for Hispanics) There are, however, noticeable differences across indicators. Simulation with ADI produces the largest relative number of NH whites (about twice as much as SVI), and PHV produces the largest relative number of African Americans (over 1.2 times as much as ADI). The differences in percentages translate into significant differences in the absolute size of the populations. For example, there is a difference of 168 thousand African Americans in ADI vulnerable tracts and PHV vulnerable tracts, a substantial number. Table 1: Policy Simulation Results Based on Statewide Rankings i Percent of Population in Vulnerable Tracts Non-Hispanic White 30% 17% 19% 15% Asian American 6% 8% 6% 9% African American 7% 8% 9% 8% Hispanic 54% 65% 63% 66% Percent of Majority Tracts Included by Indicator Non-Hispanic White 17% 5% 5% 2% Asian American 1% 1% 0% 11% African American 9% 45% 61% 38% Hispanic 41% 56% 56% 61% Hispanic, High Poverty 71% 96% 79% 97% The bottom panel in Table 1 reports the probability that majority minority neighborhoods are classified as being vulnerable. Relatively, there are few Asian American tracts in designated vulnerable tracts, due to an overall higher socioeconomic and health status25, although the findings do not reveal substantial differences among Asian ethnic groups. The ADI captures very few African American neighborhoods, while the PHV captures three-in-five African neighborhoods. The latter high rate is probably due to the multiple health problems in this population, including a high prevalence of pre-existing health conditions that increase vulnerability to COVID-19. The ADI captures the fewest Hispanic neighborhoods, while the SVI captures the most. Among the other three indicators, PHV captures fewer than HPI and SVI, perhaps due to the "Hispanic Epidemiological Paradox," where this population has better health outcome relative to their low socioeconomic status.1é Table 2 is based on designating the vulnerable neighborhoods by separate ranking for tracts within each county. Many but not all of the results are similar to those based on statewide rankings. As expected, African Americans and Hispanics comprise a larger percent of the population in high vulnerability tracts because these two groups are relatively more disadvantaged. (Again, for comparison, the following are the percentages for all tracts in the study: 27% for non-Hispanic Whites, 14% for Asian Americans, 6% for African American and 39% for Hispanics.) Compared with the statewide analysis, the differences across indicators in the percentages for all four ethnoracial groups are smaller. For example, African Americans as a percent of the population in vulnerable places range from 8.5% for PHV to 8.0% for SVI based on county rankings. One notable outcome is that county-based rankings include more Asian Americans than statewide rankings. Table 2: Policy Simulation Results Based on Within County Rankings i i yn Percent of Population in Vulnerable Tracts Non-Hispanic White 20% 19% 18% 18% Asian American 11% 11% 9% 11% African American 8% 8% 8% 8% Hispanic 58% 60% 62% 60% Percent of Majority Tracts Included by Indicator Non-Hispanic White 9% 1% 5% 5% Asian American 19% 17% 12% 21% African American 41% 44% 65% 33% Hispanic 47% 49% 53% 52% Hispanic, High Poverty 72% 84% 59% 87% Conclusion One of the underlying motivations for this study is addressing the politically contentious prohibition against race/ethnic-based policies to rectify past injustices by "leveling the playing field." The 1996 Proposition 209 imposed severe restrictions on California, forbidding government and public institutions from using race/ethnicity (and other forms of social identities) in giving "preference" in the provision and distribution of goods and services." Californian revisited Proposition 209 in the November 2020 election, but the voters refused to overturn the prohibition. This does not mean, however, that racism cannot be acknowledged. Californians rejected the 2003 Proposition 54, what would have prohibited the collection of race/ethnic data. One of the strongest arguments against the initiative was the critical need for race/ethnic data to track and analyze health outcomes. The legal constraint creates a policy dilemma.8 It is possible to identify systematic and implicit racial and ethnic biases in California's racial-neutral policy to prioritize neighborhoods high-impact by the pandemic, but it is not permissible to use race/ethnic-conscious policy instruments to rectify the flaw. The proposed solution is to utilize indirect measurements (indicators) based on underlying mechanisms that play major roles in generating racial and ethnic disparities. The findings from the policy-oriented simulation findings provides some insight into what could be a good proxy to capture underlying racial disparities. The results show noticeable differences in the groups and places designated as being vulnerable; consequently, the choice has implications in terms of who is served and who is not along racial lines. ADI performs the worst in terms of capturing African American and Hispanic persons and neighborhoods. There are also noticeable differences among the other three indicators. PHV tends to capture more African American persons and neighborhoods, probably because this group are more likely to have pre-existing health conditions. HPI and SVI have many similar outcomes, probably due to having many common input data from the American Community Survey. Base on a strict interpretation of Proposition 209, the SVI indicator could be considered not a viable option because one of its input variables is based on race. Among the remaining three (ADI, HPI and PHV), which do not explicitly include race, the PHV is the most likely to be inclusive of people and neighborhoods of color. The study's limited assessment does not indicate which indicator is "best" by other criteria and for other policy goals beyond racial justice. An analytical better approach to evaluating the indicators' usefulness is to examine how an indicator aligns with explicit policy objectives. For example, the PHV is more useful in identifying the neighborhoods that would be more adversely impacted by new waves of coronavirus infections because of more severe health and mortality outcomes. This means that inoculating residents in high PHV neighborhoods would be a desirable preventive action. The PHV, however, is not effective in directly identifying social and economic vulnerabilities. While the other three (ADI, HPI and SVI) include these types of risks, it is difficult to fully or simply understand their relative strengths and weaknesses without identifying the most relevant type and form of vulnerability. The three indicators' strength is also their weakness. They include a very broad range of factors that enable them to rank generalized vulnerability, but this breadth also makes them difficult to interpret for specific risks (e.g., probability of spreading infection associated with the built environment, or communication and trust barriers to accepting vaccines). An alternative is developing and constructing new and customized indicators that are aligned to the specific vulnerability characteristics associated with the pandemic and policy goals. 12 UCLA Center for Neighborhood Knowledge Appendix Ong, Paul M; Gonzalez, Silvia; Pech, Chhandara; Hernandez, Kassandra; Dominguez-Villegas, Rodrigo, December 15, 2020. "Disparities in the Distribution of Paycheck Protection Program Funds". UCLA Center for Neighborhood Knowledge and UCLA Latino Policy and Politics Initiative. 2020. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/12/PPP-Neighborhood- Analysis-Report_10.13.2020.pdf Ong, Paul M. December 9, 2020. "COVID-19 and the Digital Divide in Virtual Learning, Fall 2020." UCLA Center for Neighborhood Knowledge. 2020. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/12/DigitalDivide_phase2.pdf Ong, Paul M; Pech, Chhandara; Gutierrez, Nataly Rios; Mays, Vickie M, November 23, 2020. "COVID-19 Vulnerability Indicators: California Data for Equity in Public Health Decision- Making". UCLA Center for Neighborhood Knowledge and BRITE Center for Science, Research, and Policy, 2020. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/11/ CNK_CA COVID19 Medical Vulnerability 11 23 20 Final.pdf Ray, Rosalie Singerman; Ong, Paul M. November 23, 2020. "Unequal Access to Remote Work During the COVID-19 Pandemic" UCLA Center for Neighborhood Knowledge and Institute of Transportation Studies, 2020. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/12/RemoteWork_v02.pdf Ong, Paul M; Pech, Chhandara; Gutierrez, Nataly Rios; Mays, Vickie M., November 19, 2020. "Los Angeles Neighborhoods and COVID-19 Medical Vulnerability Indicators: A Local Data Model for Equity in Public Health Decision-Making". 2020. https://drive.google.com/file/d/1rRFEmrfFd5 Td iVTDebSf5tYW__26 av/view?usp=sharing Larson, Tom; Ong, Paul M, Mar, Don, and Peoples, James H, Jr. November 11, 2020. "Inequality and COVID-19 Food Insecurity". 2020. https://drive.google.com/file/d/1-fZfjjxO8aR7b8iobhmxlyllgdT8KmiKb/view Ong, Paul, Comandom, Andre, DiRago, Nicholas, Harper, Lauren. October 30, 2020. "COVID-19 Impacts on Minority Businesses and Systemic Inequality". 2020. https://drive.google.com/file/d/1gO0YT6c-Mcpx3NoknUEFfKHg4gCkvzd50/view Peoples, James H., Jr., Ong, Paul M, Mar, Don, Larson, Tom. October 28, 2020. "COVID-19 and the Digital Divide in Virtual Learning". 2020. https://drive.google.com/file/d/130CZ4|nfmYeYNPnEIAfC8SY_8G9WINLX/view?usp=sharing Ong, Paul M, Pech, Chhandara, and Potter, Megan. October 1, 2020. "California Neighborhoods and COVID-19 Vulnerabilities". 2020. https://drive.google.com/file/d/1T5U9hOvoHg506Rmvyi--I2pn4EHGhRu3L/view?ts=5f7671a8 Ong, Paul M, Mar, Don, Larson, Tom, and Peoples, James H, Jr. September 9, 2020. "Inequality and COVID-19 Job Displacement". UCLA Center for Neighborhood Knowledge and Ong & Associates, 2020. https://drive.google.com/file/d/1JEOkWRggo8zvYOdP5ribsviDimyLPxg7/view?usp=sharing Wong, Karna, Ong, Paul M, and Gonzalez, Silvia R. August 27, 2020. "Systemic Racial Inequality and the COVID-19 Homeowner Crisis". UCLA Center for Neighborhood Knowledge, UCLA Ziman Center for Real Estate, and Ong & Associates, 2020. https://www.anderson.ucla.edu/documents/areas/ctr/ziman/Systemic-Racial-Inequality-and- COVID-19-Homeowner-Crisis Wong Ong Gonzalez.pdf Ong, Paul M. August 7, 2020. "Systemic Racial Inequality and the COVID-19 Renter Crisis". Technical Report, UCLA Institute on Inequality and Democracy, Ong & Associates, and UCLA Center for Neighborhood Knowledge, 2020. https://drive.google.com/file/d/1jSRYH2EEmWeP12Db0QKDjjW8sD3L45u6/view Ong, Paul M and Ong, Jonathan. August 18, 2020. "Persistent Shortfalls and Racial/Class Disparities". Technical Report, UCLA Asian American Studies Center, UCLA Center for Neighborhood Knowledge, and Ong & Associates, 2020. http://www.aasc.ucla.edu/resources/policyreports/COVID19_CensusUpdate_ CNK_AASC.pdf Mar, Donald; Ong, Paul M. July 20, 2020. "COVID-19's Employment Disruption to Asian Americans" Technical Report, Ong & Associates, UCLA Center for Neighborhood Knowledge, and UCLA Asian American Studies Center, 2020. http://www.aasc.ucla.edu/resources/policyreports/COVID19_Employment_CNK- AASC_072020.pdf McKeever James; Ong, Jonathan; Ong, Paul M. June 25, 2020. "Economic Impact of the COVID-19, Pandemic in Riverside County, Unemployment Insurance Coverage and Regional Inequality." June 2020, Economy White Paper Series, UC Riverside Center for Economic Forecasting and Development and UCLA Center for Neighborhood Knowledge. https:// knowledge.luskin.ucla.edu/wp-content/uploads/2020/06/UCR-Economy-White- Paper COVID Ul.pdf Ong, Paul M and Ong, Jonathan. June 11, 2020. "Persistent Shortfall and Racial/Class Disparities, 2020 Census Self-Response Rate." UCLA Luskin School of Public Affairs and UCLA Center for Neighborhood Knowledge. UCLA Latino Policy & Politics Initiative and Center for Neighborhood Knowledge. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/06/ Census _ 2020 RR States 6.14.2020.pdf xOng, Paul M; Gonzalez, Silvia R; Pech, Chhandara; Diaz, Sonja; Ong, Jonathan; Ong, Elena; Aguilar, Julie. June 11, 2020. "Jobless During A Global Pandemic: The Disparate Impact of COVID-19 on Workers of Color in the World's Fifth Largest Economy." Technical Report. UCLA Latino Policy & Politics Initiative and Center for Neighborhood Knowledge. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/07/LPPI-CNK-Unemployment- Report-res-1.pdf Ong, Paul M; Gonzalez, Silvia R; Pech, Chhandara; Diaz, Sonja; Ong, Jonathan; Ong, Elena; Aguilar, Julie. May 19, 2020. "Struggling to Stay Home: How COVID-19 Shelter in Place Policies Affect Los Angeles County's Black and Latino Neighborhoods." Technical Report. UCLA Latino Policy & Politics Initiative and Center for Neighborhood Knowledge. https://latino.ucla.edu/ wp-content/uploads/2020/05/LPPI-CNK-3-Shelter-in-Place-res-1.pdf Akee, Randall; Ong, Paul M; Rodriguez-Lonebear, Desi. "US Census Response Rates on American Indian Reservations in the 2020 Census and in the 2010 Census." Technical Report. UCLA Center for Neighborhood Knowledge and American Indian Studies Center. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/05/US-Census-Response-Rates- on-American-Indian-Reservationss-051520.pdf Mar, Don; and Ong, Jonathan. "At-Risk Workers of Covid-19 by Neighborhood in the San Francisco Bay Area." Technical Report. UCLA Center for Neighborhood Knowledge. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/05/Covid-19SFBayArea.pdf Ong, Paul M; Ong, Elena; Ong, Jonathan. May 7 and May 12, 2020 "Los Angeles County 2020 Census Response Rate Falling Behind 11 Percentage Points and a Third of a Million Lower than 2010," Technical Report. UCLA Center for Neighborhood Knowledge. https://knowledge.luskin.ucla.edu/wp-content/uploads/2020/05/ Census 2020 RR_LACo final.pdf Ong, Paul M; Pech, Chhandara; Ong, Elena; Gonzalez, Silvia R; Ong, Jonathan. "Economic Impacts of the COVID-19 Crisis in Los Angeles: Identifying Renter-Vulnerable Neighborhoods," Technical Report. UCLA Center for Neighborhood Knowledge and UCLA Ziman Center for Real Estate. https://www.anderson.ucla.edu/documents/areas/ctr/ziman/UCLA- CNK_OngAssoc. LA Renter Vulnerability 4-30-20.pdf Parks, Virginia; Houston, Douglas; Ong, Paul M; Kim, Youjin B. April 24, 2020. "Economic Impacts of the COVID-19 Crisis in Orange County, California: Neighborhood Gaps in Unemployment-Insurance Coverage." Technical Report, UCLA Center for Neighborhood Knowledge and UC Irvine Urban Planning, 2020. https://socialecology.uci.edu/sites/default/ files/users/mkcruz/oc_ economic impacts of covid apr24 2020-2.pdf Ong, Paul M and Sonja Diaz. April 23, 2020 "Supporting Latino and Asian Communities During COVID-19," an opinion piece for NBC New. https://latino.ucla.edu/wp-content/uploads/2020/04/LPPI-CNK-Brief-2-with-added-notes- res.pdf Ong, Paul M; Ong, Jonathan; Ong, Elena; Carrasquillo, Andrés. April 22, 2020. "Neighborhood Inequality in Shelter-in-Place Burden: Impacts of COVID-19 in Los Angeles," Technical Report. UCLA Center for Neighborhood Knowledge and UCLA Institute on Inequality and Democracy. https://ucla.app.box.com/s/ihyb5sfqbgjrkp8jvmwiwv7bsb0u83c4 Ong, Paul M; Pech, Chhandara; Gonzalez, Silvia R; Diaz, Sonja; Ong, Jonathan; Ong, Elena. April 14, 2020. "Left Behind During a Global Pandemic: An Analysis of Los Angeles County Neighborhoods at Risk of Not Receiving Individual Stimulus Rebates Under the CARES Act," Technical Report, UCLA Latino Policy & Politics Initiative and Center for Neighborhood Knowledge. https://latino.ucla.edu/wp-content/uploads/2020/04/LPPI-CNK-Brief-2-with-added-notes- res.pdf Ong, Paul M; Pech, Chhandara; Gonzalez, Silvia R; Vasquez-Noriega, Carla. April 1, 2020. April 1, 2020. "Implications of COVID-19 on at risk workers by neighborhood in Los Angeles," Technical Report, UCLA Latino Policy & Politics Initiative and Center for Neighborhood Knowledge, 2020. https://latino.ucla.edu/wp-content/uploads/2020/04/LPPI-Implications- from-COVID-19-res2.pdf Endnotes 1 McClung N, Chamberland M, Kinlaw K, et al. The Advisory Committee on Immunization Practices' Ethical Principles for Allocating Initial Supplies of COVID-19 Vaccine - United States, 2020. MMWR Morb Mortal Wkly Rep 2020;69:1 782-1786. DOI: http://dx.doi.org/ 10.15585/mmwr.mm6947e3external icon. 2 National Academies of Sciences, Engineering, and Medicine. A Framework for Equitable Allocation of COVID-19 Vaccine. 2020. Washington, DC: The National Academies Press. doi: 10.17226/25917. 3 Michaud J, Kates J, Dolan R, Tolbert J. States Are Getting Ready to Distribute COVID-19 Vaccines. What Do Their Plans Tell Us So Far? Published: Nov 18, 2020 https://www.kff.org/ coronavirus-covid-1 9/issue-brief/states-are-getting-ready-to-distribute-covid-19-vaccines- what-do-their-plans-tell-us-so-far/. Accessed December 20, 2020. 4 California Department of Public Health, "COVID-19 Race and Ethnicity Data January 6, 2021," https://www.cdph.ca.gov/Programs/CID/DCDC/Pages/COVID-19/Race- Ethnicity.aspx. Accessed January 8, 2021. 5 Josie Huang, "Coronavirus Is Hitting Long Beach's Cambodian Community. But How Hard?" https://laist.com/2020/05/14/coronavirus-southeast-asians-cambodians-long-beach- california-covid-19.php; Rong-Gong Lin, December 31, 2020 "Coronavirus has besieged Filipino, Vietnamese Americans in Bay Area," https://www.latimes.com/california/story/ 2020-12-31/filipino-vietnamese-americans-coronavirus-silicon-valley. Accessed January 8, 2021. 6 Flanagan BE, Gregory EW, Hallisey EJ, Heitgerd JL, Lewis B. A social vulnerability index for disaster management. Journal of homeland security and emergency management. 2011 Jan 5;8(1). 7U.S. Centers for Disease Control and Prevention. Social Vulnerability Index 2018 Documentation. https://svi.cdc.gov/Documents/Data/2018_SVI_Data/ SVI2018Documentation-508.pdf. Accessed October 30, 2020. 8 University of Wisconsin School of Medicine and Public Health. "Neighborhood Atlas," https://www.neighborhoodatlas.medicine.wisc.edu/. Accessed January 8, 2021. 9 Delaney T, Dominie W, Dowling H, et al. Healthy Places Index. https://healthyplacesindex.org/wp-content/uploads/2018/07/ HPI2Documentation2018-07-08-FINAL.pdf Accessed November 1, 2020. 10 Ong PM, Pech C, Gutierrez N, Mays V. COVID-19 medical vulnerability indicators: A California data model for equity in public health state level decision-making. 2020. https:// knowledge.luskin.ucla.edu/wp-content/uploads/2020/11/ CNK_CA_COVID19_Medical_Vulnerability_11_23_20_Final.pdf. Accessed October 30, 2020. 11 UCLA Center for Health Policy Research. (2020a). "California Health Interview Survey," https://healthpolicy.ucla.edu/chis/Pages/default.aspx. Accessed November 1, 2020. 12 California Office of Environmental Health Hazard Assessment. "CalEnviroScreen 3.0, Jun 25, 2018." https://oehha.ca.gov/calenviroscreen/report/calenviroscreen-30. Accessed January 8, 2021. 13 Ong, Paul M., and Silvia R. Gonzalez. Uneven Urbanscape: Spatial Structures and Ethnoracial Inequality. Cambridge University Press, 2019. 14 This study focuses only on the major race/ethnic groups. Future analyses will examine smaller minority populations. 15 For information on health and race, see Baciu A, Negussie Y, Geller A, et al., editors. Communities in Action: Pathways to Health Equity. Washington (DC): National Academies Press (US); 2017 Jan 11. 16 Morales, Leo S., Marielena Lara, Raynard S. Kington, Robert O. Valdez, and Jose J. Escarce. "Socioeconomic, cultural, and behavioral factors affecting Hispanic health outcomes." Journal of health care for the poor and underserved 13, no. 4 (2002): 477; and Gonzalez, Silvia. "The Role of Neighborhoods and Ethnorace in Constructing Health-Related Disparities in California." PhD diss., University of California, Los Angeles, 2020. 17 Paul Ong, editor, Impacts of Affirmative Action: Policies and Consequences in California, AltaMira Press, 1999. 18 For discussion on this issue within the health arena, see Schmidt, Harald, Lawrence O. Gostin, and Michelle A. Williams. "Is it lawful and ethical to prioritize racial minorities for COVID-19 vaccines?." Jama 324, no. 20 (2020): 2023-2024. Cover Page Photo: Richard Goff (via unsplash.com) UCL yN Center for Neighborhood Knowledge https://knowledge.luskin.ucla.edu/ W KnowledgeLuskin 7 UCLACenterforNeighborhoodKnowledge knowledge@luskin.ucla.edu UCLA CENTER FOR ¢, HEALTH POLICY RESEARCH .e healthpolicy.ucla.edu https://nealthpolicy.ucla.edu/ wW UCLAchpr