The Unwinding: The Impact of Public Health Emergency Policies on Labor, Housing, and Health

Data Collection

• Highlight the need for publicly available data on pandemic social support policies.
• Support the development of a national data infrastructure with administrative linkages between social programs and health outcomes data to facilitate research on the impacts of social supports on health.
• Develop state- and local-level data sources on racial wealth gaps to identify ways to address racialized economic hardship and health inequities.
• Improve funding for longitudinal data collection to understand how policies can implement changes over time.

Study Design

• Integrate multiple kinds of study designs—including descriptive research, ethnographic studies backed by large-scale empirical data, qualitative analysis, and sophisticated causal analyses—to fully characterize public policy impacts of pandemic-era policy innovations and their relevance to future policy.
• Conduct descriptive studies to identify baseline pandemic-era trends that require causal analysis, clarify potential criteria for defining program success, and identify potential challenges in intervention implementation.
• Provide funding for long-term partnership development between researchers and social support programs to embed evaluation within programs and promote high-quality research over time.
• When designing studies, select health outcomes that are easier to disentangle from the pandemic’s direct effects, such as cardiovascular causes of death rather than anxiety, to facilitate studies of the impacts of public policy interventions.
• Continue to support and implement rapid response research programs that enable researchers to leverage “natural experiments” for policy insights.

Frameworks for Research Hypotheses

• Assess the intersecting impacts of interventions in different domains by exploring “policy bundles,” different combinations of policies, and the broader social safety net within which specific interventions are adopted.
• Understand and evaluate multidirectional support dynamics, not simply the most common forms of support (e.g., parents supporting children), throughout family and social networks.

Introduction

Christopher Whaley, PhD, Brown University

This session assessed the health economic effects of the COVID-19 pandemic and the role of public policies on the health and economic impact of the pandemic. The pandemic led to a reduction in in-person medical services, a rapid and incomplete transition to telehealth, changes to employment status, and the expansion of social support programs, as well as the rapid development of a vaccine and the immediate impacts of its distribution. Dr. Whaley emphasized the complexity of assessing the impact of public support policies during the pandemic, as the policy environment differed between states and standard evaluation approaches may not apply to policies introduced during the complex circumstances of the pandemic. Many data sources that could be used to evaluate policies could not keep up with the pace of evolution of the pandemic and saw significant delays in availability. Dr. Whaley also noted the importance of studying disparities in the health and social impacts of the pandemic in underserved communities.

Future Research Needs Arising from the COVID-Era Health Policy Experience

Kosali Simon, PhD, Indiana University

According to an analysis of data of all cause-deaths in the U.S., Canada, and United Kingdom, the United States experienced a higher COVID-19 mortality rate than similarly developed nations, likely due to an interplay of disease, policy, and individual behavior (e.g., an underprepared public health system and lower vaccination rates). The start of the pandemic led to a stark increase in public health risks and many policy changes were made during the PHE. The end of the PHE was only a policy change and no change in COVID-related deaths was observed at that time.

Public health researchers should examine several specific issues to help identify the impacts of policy changes during and after the pandemic. Medicaid recertification for enrolled patients was delayed until the end of the PHE and resulted in the highest enrollment count in history, but as pandemic-era protections expire, Medicaid patients are experiencing disenrollment. This change will disproportionately impact racial and ethnic minority populations, who use Medicaid coverage at twice the rates of White populations and experienced higher rates of mortality due to COVID-19, as evidenced by CDC data showing a significant decrease in life expectancy from 2019 to 2021 among racial and ethnic minority groups. Although little research has been published on the health outcomes of Medicaid disenrollment, the expected number of disenrollments suggests a future health crisis. The physical and mental health of children was also impacted significantly by shifts to online learning and reduced access to attention-deficit/hyperactivity disorder (ADHD) medication and diagnosis, which contributed to increased suicide rates among teenagers. Deaths due to opioids also increased rapidly during the pandemic; recent research shows that years of life lost due to opioids in 2020 was comparable to the years of life lost from COVID-19. Despite the prominence of data on opioids, little data exists on fentanyl supply, the relationship between the COVID pandemic and opioid crisis, and the role of COVID in behavioral health, mental health, and alcohol and drug use. Researchers need to complete reviews of pandemic-era research to identify lessons learned and assess the new public health equilibrium. Dr. Simon also emphasized the importance of improving access to data on public health policies, economic policies, the threat of COVID-19, and individual responses to the pandemic to inform future crisis responses. Since the end of the PHE, more rapid data sources have become available, but are challenging to assess and access, particularly to ensure long-term access needed for research.

The Sudden Rise (and Fall) of Telehealth: Where Do We Go from Here?

Ateev Mehrotra, PhD, Harvard Medical School

To compensate for an estimated 60 percent reduction in outpatient visits at the start of the pandemic compared to the previous 4 years, providers and policymakers temporarily changed telehealth policies to make them more readily available and affordable. Telehealth is defined as care delivered remotely through audio or video mediums, including remote patient monitoring and electronic consultations among health providers. Based on rates of telemedicine visits per week, these changes resulted in a substantial increase in telehealth visits (which has begun to decline since early 2021). Whereas the benefits of telehealth have been widely accepted, some concerns with this care model have also emerged. The availability of telehealth may contribute to unnecessary or low-value care and result in an increase in health care spending. Research is thus ongoing into whether telehealth is beneficial overall: studies have found that telehealth treatment for strokes and high blood pressure has resulted in significant improvements in care, but only small improvements in adherence to medication and little change in the quality of treatment for mental health. The increased reliance on telehealth may also exacerbate existing disparities in the U.S. healthcare system for several reasons. Access to telehealth depends on a high-quality internet connection, which may not be available in isolated and marginalized communities, and ethnic and racial minority groups and rural populations may be less likely to utilize telehealth options. In addition, telehealth practices may not be targeting the patients that need care the most or whose conditions are poorly managed. Use rate data from groups of patients with varying severity of symptoms showed that those with worse symptoms received remote patient monitoring at a similar rate to healthier individuals.

Dr. Mehrotra emphasized that public health researchers need to identify potential regulatory and payment policies that can promote the accessibility and quality improvements of telehealth while mitigating inequitable use and increased spending. Research on the efficacy of telehealth is crucial, as many temporary policies implemented during the pandemic are set to expire in December 2024.

Damages from the COVID-19 Pandemic and Health-Related Social Support Policies

Maria Polyakova, PhD, Stanford University

The health and economic shocks of the COVID-19 pandemic, and the effects of support systems, had varied impacts across different populations in the United States. By examining population-level excess death and missing jobs data, Dr. Polyakova found that populations that experienced high mortality rates were also among the groups that experienced significant economic damages, such as Black and Hispanic Americans, those with less education, and people employed in the food service, personal care, construction, and maintenance industries.

State policies during the pandemic often involved insurance-related changes directed toward the populations most impacted by COVID-related shocks. These changes included expanded Medicaid enrollment and relaxed enrollee recertification requirements, special enrollment periods, and suspended payments on Child Health Insurance Program (CHIP) premiums. Future research should assess the impact of these policies on the mortality gap between higher and lower income families. Another area for research is the Families First Coronavirus Protection Act (FFCPA), which provided additional funding to states that maintained a certain level of Medicaid enrollment. FFCPA resulted in more than 30 percent growth in Medicaid enrollment from 2020 to 2023, according to monthly Medicaid enrollment data, but research on the health impacts of this expansion is limited. Increased insurance rates have shown mixed effects on health outcomes, with some research finding mortality benefits of Medicaid expansion for adults under 65 years old. The unwinding of Medicaid expansion is expected to result in more uninsured people under 65 and higher rates of churn in Medicaid. Future research should explore the effects of both loss of insurance and Medicaid churn on individual health outcomes. These effects will likely vary significantly across racial, ethnic, and socioeconomic groups.

Panel 1 Discussion

Challenges to Data Collection during the Pandemic

Volatile research environments, such as the COVID-19 pandemic, can pose significant barriers to timely data collection. Standard empirical approaches and frameworks may be incapable of determining the specific impacts of social support policies, particularly when multiple programs are introduced at the same time. Multiple panel members noted that the implementation of many COVID-19 social support policies simultaneously created significant endogeneity, leading research on the impacts of specific policies to be difficult. In addition, a great deal of pandemic data will not be available for analysis for several years. Researchers can review previous literature on responses to historical social support policies to inform estimates of the impacts of pandemic-era policies, such as the likely impacts of increasing unemployment insurance or Medicaid coverage by specific amounts.

Researchers have adapted to the challenges of data collection during the pandemic by using novel datasets, such as Electronic Health Record (EHR) and cell phone data. U.S. researchers have typically used excess deaths as a primary measure of mortality due to COVID-19, whereas other countries with more developed EHR systems have access to more detailed information on diagnoses. U.S. health records would benefit from the inclusion of data that tracks timing of diagnoses, testing thresholds, and behavioral decisions that influence death and mortality data.

The Impact of the Pandemic on Public Trust in Healthcare

The development and implementation of COVID-19 policies, particularly related to vaccines, were subject to an unprecedented level of public attention and scrutiny. The amplified awareness of the process of vaccine development and utilization contributed to a significant rise in debate and skepticism over both vaccination and other public health policies. Yet the development and distribution of COVID-19 vaccines was largely successful, despite some notable areas in which the vaccine distribution campaign has received criticism. Researchers and bioethicists identified racial, ethnic, demographic, and socioeconomic populations for priority vaccine distribution, including older adults. However, this multifaceted approach to prioritized vaccine distribution was met with a public demand to focus primarily on older adults. Prioritizing only older adults would have disproportionately focused on White and Asian populations, which tend to be older. Researchers could best prepare for future pandemics by examining the technological and medical successes of the COVID vaccine campaign and public health response rather than the political or social reaction. Vaccination could also be promoted by emphasizing the spillover benefits of receiving the vaccine, including protecting people’s families and close circles and protecting against future health risks. Public health researchers should aim to identify the most logical order in which to prioritize certain populations based on race, ethnicity, wealth, or age as well as which institutions or entities would be best equipped to lead policy efforts such as vaccines.

Racial and Ethnic Disparities in Medicaid Disenrollment

Members of the panel agreed on the importance of examining the mechanisms through which racial and ethnic disparities impact Medicaid utilization among racial and ethnic minority communities. Researchers could assess the registration, enrollment, and administrative aspects of Medicaid to identify areas in which policymakers could implement changes to alleviate racial and ethnic disparities.

The Impact of Grief and Emotional Stress on Health

COVID-19 mortality had a significant mental health impact on people who had lost family members, but little research has been done on the impact of losing a younger family member on older and middle-aged adults. Older adults are more socially, emotionally, and economically dependent on family members than their younger counterparts, and one in four people who died from COVID-19 was a primary caregiver, so the role of the loss of a family member on the health of older adults is crucial to examine. However, some research has been done on the impact of losing a parent on the health of children. Further research is also needed to assess the disparities in median age of death and mental health impacts of family shocks across racial and ethnic groups.

Long COVID

Whereas the early phases of the pandemic have received substantial research and policy attention, policies to counter the effects of long COVID have not received as much research focus. Long COVID can negatively impact a person’s long term physical and mental health, economic status, and ability to work, and thus more research in this area is needed.

State Licensing for Healthcare Providers

During the pandemic, many restrictions were temporarily lifted for providers caring for patients outside of the state in which they are licensed to operate. Lifting these policies not only allowed more people to receive quality health care but also permitted neighboring facilities across state lines to share staff and alleviate burden on overloaded clinics. Access to care without state border restrictions can also reduce the need for incentives for providers to operate in isolated or underserved areas, such as Medicare subsidies for physicians in rural areas, which are often expensive. Despite these successes, however, many state barriers have been reenacted since the pandemic. Data quantifying the financial impacts of removing these barriers to cross-state medical care is essential to understanding the difference in impact of state-based and national licensure systems.

Aspects of Telehealth

The COVID-19 pandemic saw a significant reduction in in-person care that was not fully offset by a corresponding increase in telehealth visits. Since the end of the PHE, in-person care has increased, and telehealth care has decreased in frequency. Little research has examined the impact of changes or disruptions to healthcare access and consumption on individuals’ health.

Other Long-Term Research Questions

More research is needed on the ways in which care is provided to older adults—including in-home care, nursing home care, and family care—to inform how policymakers respond to the effects of future crises on older people.

Day 1 Closing Remarks

Family Support and Older Adults during the COVID-19 Pandemic

Emily Wiemers, PhD, Syracuse University

Public policies can impact individuals directly based on their own characteristics (e.g., occupation, location, wealth) and indirectly through effects on others (family or non-kin). These impacts are linked to the support that flows across and within generations, including care, in-kind support, co-residence, and emotional support. These flows of support are substantial and bidirectional. For example, in total, financial support from parents to their adult children is approximated at $80 billion annually, which is about the size of EITC spending. Dr. Wiemers shared an example of how policies interact with family support: employment disruptions increase caregiving among men, and caregiving leads to employment disruptions for women. Receiving help from family members can also impact health.

Dr. Wiemers presented her team’s analyses of time-based support during the COVID-19 pandemic using data from the Health and Retirement Study (HRS). Approximately 13-15 percent of respondents reported trouble buying food and economic hardship in the HRS COVID-19 module. Analyses indicated that parent respondents who experienced economic hardship were more likely to receive financial help from their children, whereas those who reported trouble buying food were more likely to receive time help from children. These trends also showed that children were more responsive to parental needs during the pandemic than during previous years. Using the Panel Study of Income Dynamics (PSID), Dr. Wiemers’ team also evaluated downward time transfers, from grandparents to grandchildren. Analyses suggest that, compared to 2018, grandparents provided significantly more childcare hours in 2020 if their adult children were frontline workers. Future research should examine how pandemic support policies shape family support and whether associations between caregiving/receiving care and health changed during the pandemic.

Work from Home Agreements, Job Satisfaction, and Workers’ Wellbeing at Older Ages

Marco Angrisani, PhD, University of Southern California (USC)

The U.S. labor market was substantially reshaped by the COVID-19 pandemic, leading to an unprecedented social experiment in working from home (WFH) and subsequent changes in workers’ expectations of “normal” schedules and working arrangements. Dr. Angrisani’s team measured the workforce in a rapidly evolving labor market using data from the longitudinal Understanding America Study (UAS), which is a probability-based online panel managed by USC. Since the onset of the COVID-19 pandemic, the UAS has begun collecting data on the ability to WFH and other relevant parameters in a representative sample of 6,000 individuals.

Overall, approximately half of respondents can WFH, with female workers and workers with college or advanced degrees slightly more likely to have WFH capabilities compared to male workers and those with a high school diploma. The pandemic shift towards WFH most significantly impacted younger workers, who began WFH more regularly, whereas some older adults were less affected—likely because they were already electing to use WFH policies prior to the pandemic or because they are more likely to work in occupations that are less amenable to WFH. However, approximately one-third of older workers would like to WFH more often than they do. Mismatches in WFH preferences caused a dramatic drop in job satisfaction, relative to individuals with matched WFH preferences. Other factors, such as the presence of back problems and caregiver capabilities, also interact with these trends: for example, individuals took more sick leave if they had unmatched WFH preferences and back problems, compared to those without those characteristics. In addition, work-life balance is significantly worse when WFH preferences are unmatched. Overall, accommodating WFH preferences can help improve job satisfaction and work-life balance; however, researchers must consider that enhancing WFH capabilities may also widen labor inequalities in populations unable to WFH.

Sick Leave Policy in the United States: Consequences and Trends

Nicolas Ziebarth, PhD, ZEW Mannheim, University of Mannheim, and Cornell University

Nearly all Organisation for Economic Co-operation and Development (OECD) countries provide universal access to paid sick leave, but the United States does not. During the COVID-19 pandemic, the Bipartisan Families First Coronavirus Response Act (FFCRA) was passed, providing 2 weeks of emergency sick leave. This temporary change was accompanied by a longer-term shift in employer-based sick leave: prior to 2014, job-provided short-term sick leave was offered by less than 65 percent of employers, but between 2016 and 2022, that percentage dramatically increased to nearly 80 percent. Sick-leave access is more likely to be provided for full-time employment and larger firms, compared to part-time work and smaller organizations. Most U.S. cities and states have mandated paid sick leave, but these mandates vary in comprehensiveness. Most laws enable an employee to earn 1 hour of paid sick leave per 30-40 hours works, accrue up to 7 days per year, and carry over any unused days.

Dr. Ziebarth’s team uses difference-in-differences models and event studies to estimate causal factors that impact influenza-like-illness (ILI) rates. In one study, they modeled how mandating paid sick leave can reduce ILI transmission rates by 11 percent in the first year, with further reductions in ILI rates over time. Theoretical and empirical evidence suggest that paid sick leave incentivizes contagious employees to stay home and reduces infections. For example, Dr. Ziebarth’s team has estimated that FFCRA-enabled emergency sick leave prevented 15,000 new infections per day. Further, states where workers did not gain emergency sick leave experienced significantly higher rates of COVID-19 infections. Currently, 23 percent of U.S. jobs still have no paid sick leave and most of these jobs are low-wage private sector positions in small firms, particularly in food services or accommodation industries. Dr. Ziebarth emphasized that more research is needed into the impact of paid sick leave on women, access to health care, stress, and lifecycle health.

Panel 2 Discussion

Multidirectional Support Dynamics

Panelists emphasized several research needs and gaps, including the siloing of studies of different directions of support across generations. Researchers typically study support as it flows in one direction instead of assessing multi-directional flows throughout a network. Much literature has been published on caregiving provided by parents as their children transition to adulthood, but much less has been reported on support from younger generations to older adults. Panelists highlighted that studies such as the PSID and UAS are helping to enable a family systems approach to examine flows of support throughout an individual’s life cycle and connect these disparate literatures that currently focus on only one direction of support.

Panelists also highlighted observed trends in wealth transfer, particularly that minoritized populations (Black and Hispanic) tend to transfer wealth upward (from younger to older generations), whereas White families tend to transfer wealth downward. In addition to taking a multidirectional approach to understanding all types of support, understanding the impact of these trends in wealth transfer over time is critical to assessing downstream inequities.

Labor Dynamics and Cognitive Health

Researchers need to disentangle the relationships between an individual’s occupation (which may or may not be cognitively stimulating) and their cognitive abilities later in life. Individuals may have jobs that require certain cognitive skills, but other jobs may involve tasks that enhance cognitive skills—which may have implications for later cognition. In addition, prior cognitive abilities may also explain why an individual was uniquely chosen for a position, suggesting that an individual’s cognition may influence or be influenced by labor dynamics at various points.

Social Isolation and Connectedness

Panelists discussed the impact of social connectedness on health and wellbeing during the pandemic. Interestingly, recent findings indicate that individuals who regularly experienced social isolation and mental health conditions before the pandemic experienced a smaller increase in these conditions during the pandemic than individuals who did not previously experience isolation or mental health challenges. Researchers worried as the pandemic began that the previously vulnerable population would experience greater risk, but these results suggest that more concern should fall on those not experiencing regular social isolation. Overall, panelists underscored that the impact of social isolation may affect individuals differently based on their personalities. Other work has suggested that mismatched WFH preferences–in either direction–do not impact risk for mental health conditions.

Addressing Labor Inequalities

Throughout the session, panelists emphasized several inequalities that result from existing and previous policies. One factor that may address some of these inequalities is schedule flexibility. Allowing workers to arrange their schedule in a way that provides them sufficient time for job-specific tasks, caregiving, and other tasks would help many individuals juggling responsibilities. Panelists also emphasized the importance of paid leave mandates and recommended further study and data collection on paid sick leave and its potential long-term impacts. Another topic that requires study is how quality of work is impacted for individuals who work in-person when some co-workers only WFH.

Panel 3: What Can We Learn from Eviction Moratoria and Emergency Rental Assistance Programs

Introduction

Andrew Fenelon, PhD, University of Minnesota

Dr. Fenelon introduced a panel on the impacts of major policy responses to housing insecurity during the pandemic. He noted that the pandemic affected housing economics both directly, for example through rent payments, and indirectly, as the use of homes for work and childcare increased demand for space, which in turn affected housing markets. Dr. Fenelon noted the stakes of these effects, observing that losing a home is not just an effect but also a cause of economic precarity, with implications for long-term health and wellbeing.

Emergency Rental Assistance, Housing Stability, and Health Outcomes

Vincent Reina, PhD, Housing Initiative at Penn (HIP), University of Pennsylvania

Dr. Reina summarized six research papers, which he has co-authored with colleagues at the University of Pennsylvania, that establish a framework for studying the health impacts of emergency rental assistance (ERA) programs and explore some of those impacts.

1. A paper on program design documents the many dimensions on which ERA programs vary , such as length of support, amount of funds, and coverage of future versus past rental payments.
2. Two papers on program process use national survey data to assess (1) the administrative burden of ERA programs (PDF, 371K), including the burdens of learning, compliance, psychology, and “dual uptake,” or the need for property owners to agree to accept rental assistance, and (2) the impact of procedural (process-level) and positional (individual-level) burdens on individuals’ ability to access pandemic-era ERA. The first study found that learning, regulatory compliance, and dual uptake posed obstacles for both tenants and landlords (especially non-commercial “mom and pop” landlords). The second found that tenants were also affected by interface barriers (e.g. limited language fluency), housing instability, and poor health, and the number of barriers faced was inversely related to aid approval rates.
3. A paper examines a California-based effort to reduce these process burdens and found that it increased weekly ERA approval rates and nearly tripled odds of approval for Hispanics and people living in less formal housing units.
4. Three papers examined impacts of short- and long-term ERA during the pandemic, in Philadelphia, four cities across the U.S., and California . These studies found that ERA had positive economic and health impacts (e.g., in Philadelphia, it reduced amount of rent owed, odds of borrowing money to pay rent, and stress levels) but did not affect other measures (e.g., odds of being behind on rent or ability to consume other basic goods).

These studies collectively suggest that program design, program process, and the nature of the assistance all matter for program impact. The studies also raise specific questions for future research in these areas. Findings about variability in program design raise questions about the impacts of variable depth and length of subsidies on health outcomes. Determining the optimal depth and length could help inform new housing-related interventions, including current efforts to leverage Medicaid funds to support housing stability. Studies of program process underline the need to apply to ERA programs lessons from other fields regarding process burden. Future research on such questions requires causally linking housing and health data to characterize impacts of housing-related interventions. Such research also requires long-term partnerships between researchers and housing programs that focus on supporting those programs rather than merely using them as a source of study data. Such partnerships between HIP and various state and local governments were established long before the pandemic and enabled the studies described by Dr. Reina during this presentation; funding for such partnerships should be a part of the research landscape.

Studying the Health Impact of Eviction Before and After the Pandemic

Nick Graetz, PhD, Princeton University

Housing may impact health through multiple mechanisms: instability, unaffordability, physical quality, and neighborhood opportunity. Affordability and instability are less examined mechanisms but are important to examine as, for example, high housing costs can lead families to reduce spending on healthcare, while evictions can lead to acute and chronic trauma. However, the lack of a data infrastructure for eviction and health data has complicated study of these impacts. Existing research tools, such as surveys, often do not adequately measure evictions, and there is no systematic collection of formal eviction case records or reporting of eviction filings across states; moreover, even when eviction data are available, they still require administrative linkages to health data. To address these measurement challenges and assess the connections between housing and health, Dr. Graetz’s team has worked with the U.S. Census Bureau to build a national database of formal eviction records from housing courts, which are collected, entered, and collated by the Princeton Eviction Lab, and has linked this database to Census and administrative data. A recent study analyzing these data linkages shows that from 2000-2019, an eviction filing without a judgment was associated with an increase in the risk of all-cause mortality by about 20 percent and an eviction judgment by about 40 percent. After the study team ruled out major changes in the evicted population during the pandemic, mortality rates increased by about 106 percent for renters who received an eviction filing , compared to about 25 percent for their neighbors. These trends may lead to health disparities, given underlying disparities in evictions—which Graetz’s team has found are concentrated among adults aged 20-40, likelier for families with children, and even more likely for black women living with children.

Dr. Graetz’s team is currently working to link data on federal housing assistance to its database to assess health outcomes at federally financed properties. Future research should also consider broader health outcomes than are currently possible with individual data linkages; examine owner characteristics that are often not captured in current data systems; and identify who is responding to different housing policies and why.

Rollback of Pandemic Eviction Moratoria: Population Health Implications

Kathryn Leifheit, PhD, University of California, Los Angeles

Eviction moratoria are a strong policy intervention that might protect health by reducing stress, enabling families to spend housing budget on healthcare and other necessities, mitigating household crowding that could increase COVID-19 transmission, and alleviating pressure to take work with high COVID-19 exposure. However, these health impacts had not been empirically tested. To evaluate the moratoria’s actual health impacts, Dr. Leifheit’s team compared COVID-19 cases/death across states that instituted moratoria for different lengths of time between March 1 and September 30, 2020. The average moratorium was only 10 weeks long. The team found that 16 weeks after states lifted their moratoria, their COVID-19 incidence was 2.1 times higher and their COVID-19 mortality was 5.4 times higher compared to states that retained moratoria during that 16-week period. The stronger association with death rate than incidence rate may indicate that COVID-19 cases were more severe among the evicted population, reflecting population-level health disparities. Another team explored such disparities further by stratifying COVID-19 incidence by race; in unpublished data, they show that cases rose much more among non-Hispanic Black individuals compared to non-Hispanic White individuals. A third study examined the connection of moratorium strength to renter mental distress using data from the UAS. It found that stronger protections (i.e., that blocked eviction notices and filings) were associated with lower mental stress compared to protections that blocked only displacement.

Future research should examine the impacts of the implementation and rollback of all pandemic-era eviction moratoria, other supports (e.g., right-to-counsel), and emergency rental assistance; consider these policies’ impacts on other outcomes such as food insecurity; and explore their longer-term health impacts. By studying these impacts, researchers can leverage pandemic-era innovations to inform planning for future public health and economic crises as well as non-pandemic health-related housing supports. However, these generalizations will require disentangling the impacts of multiple pandemic-era interventions and identifying which impacts were unique to the pandemic context.

Panel 3 Discussion

Panel 3 Discussion

Translating Pandemic Policy Research for Non-Pandemic Periods

The scale of the pandemic and of policy responses to it pose challenges for interpreting intervention impacts and characterizing their generalizability. Most notably, the unprecedented level of federal innovation during the pandemic and the sheer number and variety of social support programs can complicate efforts to isolate impacts of specific programs. Such impacts may be heightened in the context of extreme pandemic-era health threats or diminished in the context of complementary economic supports. One way that researchers can address the complexities of pandemic policy studies is through their selection of health outcomes. The impacts of policy interventions on outcomes such as mental health and anxiety are difficult to disentangle from the pandemic itself, given its own direct impacts on mental health. However, studies of health outcomes that are not as intertwined with the pandemic, such as cardiovascular causes of death, could identify persistent pathways through which evictions may impact health.

Moreover, the uniqueness of the COVID-19 pandemic is not just an obstacle but a research opportunity. The pandemic created an unprecedented set of natural experiments from which rigorous research methods can yield generalizable insights. For example, during the pandemic, Dr. Reina and the city of Philadelphia created a direct cash pilot program for renters called PHL Housing+ that was officially launched in 2021. Drs. Reina and Jaffee at the University of Pennsylvania are currently evaluating the program’s impact using a randomized design. Policymakers had previously been concerned programs like PHL Housing+ would undermine rather than complement rental voucher programs, but the pandemic highlighted the assistance tools needed and that policymakers can work in tandem to help various households in different ways. Studies like those presented during this panel, which leverage robust datasets and methods, can identify health outcomes associated with such programs, assess the programs’ causal impacts, and characterize mechanisms of these causal relationships. Research can also explore the nuanced ways that negative health outcomes might materialize even in the context of such supports; these nuances are suggested by a study by Dr. Reina and colleagues that identified increased moves even during periods without formal evictions. Finally, the visibility of pandemic programs can highlight areas of study that need more attention, such as the impact of administrative burden on access to social support programs.

Studying the Intersections of Multiple Interventions

The variety of social programs available to families during the pandemic—in health, labor, housing, and other domains—require careful controls in studies of specific interventions. Yet the intersecting impacts of these programs are also worth studying directly. For example, Dr. Reina and colleagues has found that the delivery mechanisms and size of broad system-wide safety nets can dilute the effects of ERA, but these effects are still visible even in the pandemic context; this combination of results underlines the importance of ERA as an intervention and suggests its impacts can differ in combination with various kinds of social supports. Similarly, Dr. Leifheit suggested that the effects of expiring eviction moratoria in a context where ERA exists is distinct from expiration without rental support, where eviction moratoria may have simply delayed eviction. Similarly, eviction moratoria might have different health implications when implemented separately from or combined with ERA. Assessing longer-term health impacts of policy interventions might also require consideration of “policy bundles” rather than of single programs.

Measuring and Relieving Administrative Burden

Administrative burden is one aspect of policy interventions that intersects across domains: this burden is compounding and cumulative for people who frequently are accessing multiple programs at once. For example, people interacting with housing courts and completing applications for ERA may lose Supplemental Nutrition Assistance Program benefits because they cannot keep up with the collective administrative burden of multiple programs. Taking this compound view of administrative burden in research can thus be helpful. Moreover, lessons learned in one domain should also be applied to others; if housing assistance is available through Medicaid waivers, for example, those waivers would ideally not mimic aspects of housing voucher programs (such as dual uptake) that are particularly burdensome.

Participants identified several process innovations that should receive further study as ways to reduce administrative burden without compromising program integrity or accountability:

1. During the pandemic, self-attestation of income level was used to qualify applicants for housing-related assistance. The city of Los Angeles recently began using this approach to approve applications for housing support, with approval followed by formal verification.
2. The Philadelphia PHL Housing+ cash assistance pilot offers another innovation by not requiring housing inspections; future research can explore the impact of this approach on housing quality.
3. Joint eligibility processes that qualify a person for multiple programs at once could help alleviate administrative burden.

Participants also described promising research approaches, including the importance of characterizing the full arc of program processes, which are the source of administrative burden. Qualitative studies could also be useful: most studies of administrative burden to date have been qualitative (Dr. Leifheit mentioned the work of Dr. Danya Keene), and a qualitative approach could define challenges faced by people who tried to access eviction moratorium protections or ERA during the pandemic and use these data as a basis for building quantitative models.

The Importance of Descriptive Policy Research

Although causal studies are critical to informing housing and other social policies, participants emphasized the importance of a range of research methods, including descriptive research. Descriptive studies have foundational importance: they can identify baseline phenomena that require causal analysis (such as Dr. Leifheit’s finding that homelessness counts were consistent from 2016 to 2022 before rising in 2023); clarify potential criteria for measuring program success (for example, Dr. Graetz’s findings that children are particularly impacted by evictions); and point to potential challenges in intervention implementation (for example, Dr. Reina’s process research identifying the impact of administrative burden). Two specific areas in which descriptive research could be useful are the life course impacts of pandemic-era policies and spillover effects of policy interventions on family members.

Well-designed causal studies with carefully applied statistical techniques (e.g., regression modeling) are also important, particularly if they consider not only the local impacts of specific policy change on a particular outcome, but also the impacts of policy bundles or cumulative administrative burden. Ethnography supported and reinforced by large-scale empirical data can also be broadly accessible and useful.

Translating Research for Policymakers

Research on policy interventions is not always readily usable by policymakers, and participants discussed several ways that researchers can improve that usability, such as adopting study designs that are likely to offer useful policy insights and making data accessible to policymakers. Dr. Reina’s team, for example, prepared a research summary that helped the city of Los Angeles determine whether to join a state ERA program.

Research Gaps: Housing Actors and Lifecourse Impacts

Participants concluded the session by identifying two specific research questions that require further exploration. Dr. Leifheit noted that children who live in states or localities with more supportive housing policies can be predicted to have better lifecourse outcomes than children who lived in states or localities with weaker policies, and suggested that this hypothesis should be examined—potentially in relation to “policy bundles” rather than only to housing policy specifically. Dr. Graetz called for efforts to collect better data on key housing sector actors—such as owners, lenders, and insurers—and related characteristics (such as ownership turnover and forms of financing) that can affect housing and health.

Closing Remarks

Dr. Erin Harrell thanked all participants and underscored the importance of the topics covered in this workshop.

Tuesday, May 6, 2024

1:00 p.m.-2:10 p.m. | Opening Remarks and Overview of Pandemic-Era Social Support Policies

• John Phillips, PhD, NIA
• Julia Raifman, PhD, Boston University
• Kriti Jain, PhD, NIA

2:10 p.m.-2:20 p.m. | Break

2:20 p.m.-5:00 p.m. | Panel 1: Changes to Healthcare Coverage and Delivery from the COVID-19 Pandemic

• Christopher Whaley, PhD, Brown University
• Kosali Simon, PhD, Indiana University
• Ateev Mehrotra, PhD, Harvard University
• Maria Polyakova, PhD, Stanford University

Monday, May 20, 2024

1:00 p.m.-1:05 p.m. | Day 2 Welcome and Overview Kriti Jain, PhD, NIA

1:05 p.m.-2:55 p.m. | Panel 2: Effects of Income and Labor Policies Enacted During the COVID-19 Pandemic

• Kathleen Mullen, PhD, University of Oregon
• Emily Wiemers, PhD, Syracuse University
• Marco Angrisani, PhD, University of Southern California
• Nicolas Ziebarth, PhD, ZEW Mannheim, University of Mannheim, and Cornell University

2:55 p.m.-3:10 p.m. | Break

3:10 p.m.-4:55 p.m. | Panel 3: What Can We Learn from Eviction Moratoria and Emergency Rental Assistance Programs?

• Andrew Fenelon, PhD, University of Minnesota
• Vincent Reina, PhD, University of Pennsylvania
• Nick Graetz, PhD, Princeton University
• Kathryn Leifheit, PhD, University of California, Los Angeles

Appendix B: Meeting Participants