Access and enrollment in safety net programs in the wake of COVID-19: A national cross-sectional survey.

The global COVID-19 pandemic is causing unprecedented job loss and financial strain. It is unclear how those most directly experiencing economic impacts may seek assistance from disparate safety net programs. To identify self-reported economic hardship and enrollment in major safety net programs before and early in the COVID-19 pandemic, we compared individuals with COVID-19 related employment or earnings reduction with other individuals. We created a set of questions related to COVID-19 economic impact that was added to a cross-sectional, nationally representative online survey of American adults (age ≥18, English-speaking) in the AmeriSpeak panel fielded from April 23-27, 2020. All analyses were weighted to account for survey non-response and known oversampling probabilities. We calculated unadjusted bivariate differences, comparing people with and without COVID-19 employment and earnings reductions with other individuals. Our study looked primarily at awareness and enrollment in seven major safety net programs before and since the pandemic (Medicaid, health insurance marketplaces/exchanges, unemployment insurance, food pantries/free meals, housing/renters assistance, SNAP, and TANF). Overall, 28.1% of all individuals experienced an employment reduction (job loss or reduced earnings). Prior to the pandemic, 39.0% of the sample was enrolled in ≥1 safety net program, and 50.0% of individuals who subsequently experienced COVID-19 employment reduction were enrolled in at least one safety net program. Those who experienced COVID-19 employment reduction versus those who did not were significantly more likely to have applied or enrolled in ≥1 program (45.9% versus 11.7%, p<0.001) and also significantly more likely to specifically have enrolled in unemployment insurance (29.4% versus 5.4%, p < .001) and SNAP (16.8% versus 2.8%, p = 0.028). The economic devastation from COVID-19 increases the importance of a robust safety net.

Introduction

The United States is at the center of the global COVID-19 pandemic, with more than one-quarter of the world’s confirmed cases and one-fifth of all deaths as of mid-August 2020 [1]. Along with surging mortality rates, COVID-19 has brought economic devastation to many American families. Beginning in March 2020, most states enacted mandatory shelter-in-place orders to reduce transmission and many businesses closed, leading to decreased economic activity and widespread layoffs. From March to April 2020, an estimated 20.5 million individuals became unemployed and the unemployment rate reached 14.7%, rebounding to 10.2% in July 2020 [2].

National surveys from early in the pandemic highlight the financial strain faced by US families. A survey conducted from April 15–20, 2020, found that 31.0% of households experienced difficulties affording basic needs or paying bills [3]. One-fifth of households experienced inadequate access to food in April 2020, with higher rates among families with young children [4]. The burden of these challenges is falling disproportionately on people who are part-time workers, have children, are younger, or are racial/ethnic minorities [5, 6]. While economic activity resumed in many states in May 2020, the resurgence of the epidemic in many parts of the U.S. in the summer prompted a new wave of business and school closures. The Census Bureau’s Pulse Household Survey documented that family hardship–including food scarcity and housing insecurity–trended upward from April 23 to July 21, 2020 (the most recently available data for the current study) [7].

Many individuals turned to safety net programs for assistance early in the pandemic. News media reports have profiled the growing demand on food pantries and social assistance programs [8, 9]. Unemployment insurance programs, which are a joint federal-state partnership to support displaced workers, are also seeing enormous demand [10]. Means-tested, federally-funded programs such as Medicaid, the Supplemental Nutrition Assistance Program (SNAP), and Temporary Assistance for Needy Families (TANF) are expected to provide some assistance for low-income individuals. However, resources for programs like TANF are limited due to their block-grant structure, income eligibility, and work requirements [11]. Assistance programs were given a small one-time boost in the $2.0 trillion Coronavirus Aid, Relief, and Economic Security (CARES) Act, which extended the duration of unemployment insurance benefits and provided $1,200 in direct payment to eligible adults and $500 to dependent children [12]. In late August 2020, with federal unemployment assistance set to expire, the US Congress was at an impasse over the scope of a potential relief measure. While Democrats were pushing for a renewal of the $600 weekly unemployment insurance payments, the Trump Administration and some Congressional Republicans resisted a more extensive relief package [13].

As Congress and the states contemplate further investments in safety net programs, there is an urgent need to identify how families affected by COVID-19 have accessed available programs before and since the pandemic and the challenges they expected to confront. Further, given the substantial economic burden of the COVID-19 pandemic on already disadvantaged groups, it is important to identify those most in need in order to better target safety net funding and program activities, including outreach and enrollment. Further, data from early in the pandemic provides an important baseline for evaluating the evolving changes in program participation and hardship among vulnerable individuals. We therefore conducted a nationally representative survey regarding the COVID-19-related hardships experienced by American adults and their use of safety net services early in the pandemic, and hypothesized that the impact of COVID-19 and enrollment in safety net programs would be greatest among those experiencing COVID-19-related employment and earnings loss.

Methods

We collected data using the AmeriSpeak Omnibus survey, a nationally representative panel of American households recruited and maintained by NORC at the University of Chicago. The Omnibus survey is a cross-sectional survey of a rotating set of the AmeriSpeak panel that is conducted every two weeks and to which researchers and other partners can contribute items. The AmeriSpeak panel is recruited using stratified, address-based sampling methods that cover approximately 97.0% of all residential addresses. The multi-stage probability sample is created using a national frame area where blocks are sampled from within defined metropolitan or rural areas. AmeriSpeak oversamples in areas with a higher concentration of young adults and minorities and engages in additional efforts to follow up with households that initially do not respond. Individuals are recruited to the panel using a combination of US mail, telephone interviews, and in-person field interviews. Households can respond to the survey by internet (including on smartphones) or by telephone interview. About 85% of the interviews are completed online and 15% are conducted over the phone. The phone option is offered to allow “net-averse” households to participate. The overall response rate for the panel is about 34.0% (American Association for Public Opinion Research [AAPOR] response rate three) [14].

For this study, our team developed the State Health Access Data Assistance Center (SHADAC) COVID-19 Safety Net Survey and contracted with NORC to add the survey questions to the survey that was in the field April 23 to April 27, 2020. We contracted with NORC to administer the survey to a target of 1,000 respondents. The study was restricted to people over age 18. The final sample included 1,007 adults. Table 1 shows the demographic characteristics of the study sample. NORC develops weights to national census benchmarks and balances by gender, age, education, race/ethnicity, and region. The weighted sample is similar to a national sample of adults: 51.4% of the sample was female, 44.8% between age 18 and 44, 37.4% non-white, 46.1% with a chronic condition, 36.2% with a high school degree or less, and 83.8% residing in metropolitan areas.

Table 1

Demographics of study sample.

	Unweighted n	Weighted %	Standard Error
Female	524	51.4	2.04
Age Group
18–29	113	18.1	1.94
30–44	268	26.7	1.78
45–59	251	24.5	1.66
60+	375	30.7	1.75
Non-white	362	37.4	2.00
Any chronic condition	492	46.1	2.05
Education
No HS diploma	36	8.8	1.53
HS graduate or equivalent	126	27.5	2.13
Some college	293	28.5	1.72
BA or above	552	35.3	1.70
Resides in a metro area	885	83.8	1.65

Note: Weighting performed using survey weight created by NORC to approximate to national proportions. Sample size = 1,007 individuals.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

The current study analyzes unemployment reduction, economic burden, and use of government safety net and assistance programs. The main comparison of interest is between individuals who experienced either job loss or loss of income from COVID-19 (which we call “employment reduction”) versus those who did not (see Table 2 for categories). We examined whether individuals in both groups were aware of seven major safety net programs (Medicaid, health insurance marketplaces, unemployment insurance, food pantries/free meals, housing/renters assistance, SNAP, and TANF). We further examined whether individuals said that they were enrolled before the pandemic in each of these programs and whether they had enrolled or applied since the pandemic. We also asked individuals how they intended to spend their $1,200 stimulus checks to assess spending priorities. Finally, we asked individuals to rate their confidence in their ability to pay for several basic needs and expenses over the next four weeks from date of the survey. The items that were used in the survey were developed by our team for the purpose of this study; items were not piloted before being used in the study.

Table 2

Experience of employment or earnings loss related to COVID-19.

Category		Percent	Standard Error
Coronavirus Employment or Earnings Loss		28.1%	1.9%
	I have had my work hours cut due to the coronavirus	12.3%	1.3%
	I have lost my job due to the coronavirus	10.6%	1.5%
	I have had my pay cut due to the coronavirus	5.3%	0.8%
	I have retired from work due to the coronavirus	1.3%	0.5%
	I am on paid leave because my employer closed due to the coronavirus	2.4%	0.5%
No Coronavirus Employment or Earnings Loss		71.0%	1.9%
	I was not employed at the onset of the coronavirus	28.3%	1.8%
	I am working from home due to the coronavirus	17.3%	1.4%
	I have gotten a new job because of the coronavirus	0.3%	0.3%
	I am working more hours due to the coronavirus	7.7%	1.2%
	The coronavirus has not affected my job	20.4%	1.7%

Note: Individuals can identify more than one factor for employment status so categories sum to greater than 100%.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

Each of our main outcomes was dichotomized. Survey weights were used in all analyses to account for known differences in sampling probability and non-response. In bivariate analysis, we compared individuals with and without COVID-19-related employment or earnings loss on their safety net awareness and participation, plans for spending stimulus checks, and confidence in ability to pay for basic needs. We calculated two-sided t-tests for differences in means between the groups. Because we were interested in identifying the disproportionate changes in program participation among those experiencing employment reduction, we fit a regression model for program participation that estimates the average change since the pandemic, the baseline rate for those without employment reduction, and an interaction term. This interaction term is analogous to a difference-in-differences coefficient, representing the change in program participation since the pandemic for those with employment reduction versus those who without employment reduction. The study was determined exempt by the University of Minnesota Institutional Review Board (IRB). [Study data will be archived at YYY DOI number ZZZ after paper acceptance].

Results

Overall, 28.1% of respondents said that they experienced an employment or earning loss due to the coronavirus, with 10.6% specifically reporting losing a job, 12.3% reporting losing work hours, and 5.3% reporting a pay cut (Table 2).

Respondents’ awareness and enrollment in seven major safety net programs before and after the pandemic is recorded in Table 3. Overall, a majority reported that they were aware of each program, with the highest awareness for SNAP (95.6%), Medicaid (91.9%), and food pantries (89.2%) and the lowest awareness reported for housing/renters assistance (69.8%) and TANF (63.8%). Correspondingly, overall program participation was highest prior to COVID-19 in Medicaid (21.9%) and SNAP (20.1%). Prior to the pandemic, 39.0% of the sample was enrolled in at least one safety net program. Overall, there were significant increases in the percent of people reporting that they had applied for or enrolled in most safety net programs since the pandemic. The overall percentage reporting at least one program increased by 15.29 percentage points (p<0.0001), with the largest reported changes for unemployment insurance (7.87 percentage points, p<0.0001) and SNAP (4.27 percentage points, p<0.0001).

Table 3

Awareness and enrollment in safety net programs before and since pandemic.

Program	Awareness of program	Enrollment in the Program
		Prior to the Pandemic	Since the Pandemic	Change in Enrollment	P-Value for Change
At Least One Safety Net Program	98.00%	39.00%	46.72%	15.29	p<0.001
Medicaid	91.90%	21.90%	23.61%	1.70	p<0.001
Health insurance exchanges	71.20%	11.30%	13.11%	1.80	0.001
Unemployment insurance	77.70%	9.20%	17.05%	7.87	p<0.001
Food pantry/free meals	89.20%	11.50%	14.86%	3.33	p<0.001
Housing/renters assistance	69.80%	5.70%	6.72%	1.07	0.009
SNAP	95.60%	20.10%	24.32%	4.27	p<0.001
TANF	63.80%	1.70%	2.55%	0.90	0.009

Note: P-value represents a t-test for the change in enrollment prior to the pandemic and since the pandemic. Change in enrollment is in percentage points.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

Table 4 disaggregates the safety net program participation data before and since the pandemic between people who did versus did not experience employment reduction. Participation in at least one safety net program prior to the pandemic was higher among people who subsequently experienced COVID-19 employment reduction than those not experiencing employment reduction (50.0% versus 37.7%). Individuals experiencing employment reduction were more likely to have participated in the health insurance exchanges (23.0% versus 10.0%) and SNAP (35.3% versus 18.3%) before the pandemic than those who did not. The difference-in-differences coefficient, which identifies the relative difference in program participation since the pandemic between the two groups shows that those in the employment reduction group were significantly more likely to have enrolled in at least one program since the pandemic (22.2 percentage point increase, p = 0.038). The increase was particularly notable for increased participation in unemployment insurance (24.0 percentage point increase, p = 0.0009).

Table 4

Comparing changes in program participation among individuals who experienced employment reduction versus those who did not.

	Prior to the Pandemic		Since the Pandemic		Change in Enrollment (Difference)*		Diff-in-Diff*	p-value for Diff-in-diff
	ER	NER	ER	NER	ER	NER
At Least One Safety Net Program	50.00%	37.70%	77.50%	43.08%	27.57	5.37	22.20	0.038
Medicaid	29.70%	21.00%	36.31%	22.11%	6.60	1.10	5.50	0.596
Health insurance exchanges	23.00%	10.00%	30.62%	11.10%	7.60	1.10	6.50	0.505
Unemployment insurance	5.30%	9.60%	34.65%	14.99%	29.40	5.40	24.00	p<0.001
Food pantry/free meals	12.20%	11.50%	16.54%	14.67%	4.40	3.20	1.20	0.882
Housing/renters assistance	5.20%	5.70%	10.37%	6.30%	5.10	0.60	4.50	0.335
SNAP	35.30%	18.30%	52.07%	21.06%	16.80	2.80	14.00	0.201
TANF	3.10%	1.50%	5.83%	2.16%	2.70	0.70	2.00	0.671

Notes: “ER” = COVID-19 related employment reduction, “NER” = no COVID-19 related employment reduction. Diff-in-diff represents the change in enrollment since the pandemic for ER group relative to the NER and is estimated from a regression model that includes an interaction between “post” pandemic and being in the ER group. P-value for diff-in-diff is the p-value associated with that interaction term.

*Unit reported is a percentage point change.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

The highest priority for stimulus check spending among the provided categories was mortgage/rent (24.2%), followed by utilities (17.6%), debt and loans (16.3%), food (13.6%), savings or investment (10.3%), medical care and insurance (3.7%), and donations (2.8%) (Table 5). Compared to those with no employment reduction, people experiencing COVID-19 employment reduction were significantly more likely to plan using the checks for mortgage/rent (47.1% versus 21.6%, p = 0.003) and less likely to use it for debt (5.8% versus 17.5%, p = 0.003), donations (0.0% versus 2.0%, p<0.001), and medical expenses (0.0% versus 4.1%, p<0.001).

Table 5

Priority for stimulus check spending.

	Everyone	COVID Job Loss	No COVID Job Loss	P-Value
Mortgage or rent	24.2%	47.1%	21.6%	0.003
Utilities (electricity, water, heat, gas, internet, etc.)	17.6%	17.2%	17.6%	0.937
Food for myself/family	13.6%	12.9%	13.6%	0.888
Credit card debt, car payments, student loans	16.3%	5.8%	17.5%	0.003
Medical care or insurance premiums	3.7%	0.0%	4.1%	p<0.001
Savings or Investment	10.3%	9.1%	10.4%	0.873
Donation	2.8%	0.0%	3.1%	p<0.001

Note: We combined response categories for paying off credit card debt, making a car payment, and paying off student loans and we combined response categories for paying for medical care already received, needed medical care, and insurance premiums.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

People experiencing COVID-19-related employment reduction were significantly less likely to report being confident in their ability to pay all categories of basic expenses (Table 6). Overall, 33.5% of the sample was not confident in their ability to pay for at least one type of expense, with 69.9% of those with COVID-19 employment reduction reporting a lack of confidence in ability to pay expenses versus 28.8% of those with no COVID-19 employment reduction (p<0.001). Our analysis found significant differences between the two groups (employment reduction versus no employment reduction) in every category of expense. Confidence was lowest for ability to pay debt (55.8% versus 17.2%, p<0.001), medical care or insurance premiums (45.3% versus 15.5%, p<0.001), and mortgage or rent (35.8% versus 9.7%, p = 0.002).

Table 6

Not confident in ability to pay for basic needs over next four weeks.

Program	Everyone	COVID Job Loss	No COVID Job Loss	P-Value
Not confident in at least one of the options below	33.5%	69.9%	28.8%	p<0.001
Mortgage or rent	12.8%	35.8%	9.7%	0.002
Utilities (electricity, water, heat, gas, internet, etc.)	9.3%	34.8%	6.4%	p<0.001
Food for myself/family	6.9%	32.6%	3.9%	p<0.001
Credit card debt, car payments, student loans	21.2%	55.8%	17.2%	p<0.001
Medical care or insurance premiums	18.7%	45.3%	15.5%	p<0.001

Note: Table reports on percentages responding that they were either “not at all” or “not very” confident in their ability to pay this expense. We combined response categories for paying off credit card debt, making car payment, and paying off student loans and we combined response categories for paying for medical care already received, needed medical care, and insurance premiums.

Source: Authors’ analysis of the April 2020 SHADAC COVID-19 survey.

Discussion

This study examined the economic challenges confronting American adults in mid-April 2020, with a focus on individuals experiencing employment or earning loss related to the COVID-19 pandemic. Overall, our study findings highlight the precarity of individuals who were most affected early in the COVID-19 pandemic. While majorities of the sample were aware of safety net programs, awareness was lowest for TANF and the health insurance exchanges. We found that 45.9% of individuals experiencing COVID-19 job loss had enrolled or applied for a safety net program since the pandemic, with the highest application rates reported for unemployment insurance and SNAP. A majority of those experiencing COVID-19-related employment or earnings loss expressed concern about paying for basic needs in the month, especially the ability to pay off existing debt. Importantly, the group that experienced COVID-19-related employment reductions also tended to be the group more likely to have accessed safety net programs before the pandemic, suggesting their heightened vulnerability.

Our findings underscore concerns that have been raised about the sustainability of existing safety net programs in attempting to deal with this extraordinary public health and economic challenge to the U.S. population [15]. They also reveal gaps in knowledge of key safety net programs that could help provide continuity of insurance or benefits to those most in need of such programs. For example, roughly one-quarter of respondents did not have awareness of unemployment insurance; however, these knowledge gaps could be remedied through efforts to increase education and outreach to those who experience job loss. Furthermore, individuals who may have a general awareness of programs may not necessarily know whether they might be eligible. For example, not all workers are eligible for unemployment insurance. To qualify, individuals must acquire a sufficient work history, and the program often excludes certain types of workers (e.g., “gig” workers such as Uber drivers) [10]. Most other safety net programs include large exclusions that limit their reach to vulnerable individuals.

The stability of food assistance programs is of particular concern. The rising levels of food insecurity in the pandemic have been accompanied by a food production crisis. Early in the pandemic, some farmers were resorting to destroying agricultural products they were unable to sell [16]. Expanding food assistance programs could help address food insecurity and the food production crisis. Awareness of SNAP and food pantries were relatively high in our sample. Recent enrollment in SNAP was particularly high, especially among people with COVID-19 employment or earnings loss. However, challenges loom for SNAP enrollees. COVID-19 has delayed, but not eliminated, a recent push by the Trump Administration to increase the stringency of work requirements for SNAP enrollees. While planned work requirements were suspended by the CARES Act, these work requirements would be reinstated as soon as the official national emergency expires, but likely long before the economy sufficiently rebounds. Nearly 700,000 Americans are set to lose benefits under the current requirements, and this number is likely to grow with rising unemployment [17].

The affordability of housing is another concern. Almost half of all individuals experiencing COVID-19 employment or earnings loss said that they would spend their $1,200 stimulus checks on rent or mortgage. However, this monetary infusion offered only small and temporary relief, as median rent in the U.S. is around $1,000 per month [18]. Most states opted to suspend evictions in the context of state-declared COVID-19 emergencies, but the scope of these protections varied widely across the states [19]. States began lifting these restrictions in May 2020, leading to a resumption in eviction [20]. The CARES Act did include six months of relief from eviction for individuals paying federal mortgages and provided limited funds for community block grants and the federal Housing and Urban Development (HUD) Administration [21]. However, these provisions were unlikely to reach many of the most precarious families, including renters not living in subsidized housing.

As millions have lost their employer-sponsored health insurance it is important to consider the potential impact on public insurance programs. Programs created or expanded under the Affordable Care Act (ACA)—Medicaid and the health insurance exchanges—likely absorbed some of the new demand. In the 37 states (including the District of Columbia) that expanded Medicaid, the program now covers most individuals up to 138% of the Federal Poverty Level (FPL), while the exchanges provide slide-scale subsidies to individuals with household incomes between 100% and 400% FPL. Further, limited emergency Medicaid coverage can be extended to uninsured individuals for COVID-19-related care (including in states that have not expanded Medicaid) [22]. In our sample, 6.6% of people experiencing COVID-19 unemployment reduction indicated they had applied or enrolled in Medicaid since the pandemic and 7.6% had said the same about the insurance exchanges.

The increased enrollment in public programs could have important implications for state budgets and delivery systems. Early in the pandemic, state Medicaid programs began projecting that the weakened economy would increase program enrollment and total spending [23]. Testing and treatment related to COVID-19 could also contribute to rising Medicaid spending, though this spending could be offset by other forms of medical care that decreased after the pandemic, such as preventive office visits. State programs have a variety of policy options for easing transitions of new members into Medicaid, reducing churn, and simplifying enrollment. A recent analysis suggests that Medicaid enrollment increased early in the pandemic, although these increases were not correlated with enrollment changes in those states [24]. Similarly, some states have used their exchanges to provide special COVID-19 open enrollment period, with eligibility open to the currently uninsured due to the pandemic [25]. For those relying on the federal marketplace (healthcare.gov), no such special COVID-19 open enrollment period has thus far been implemented. Both the federal and state exchanges provide for the possibilities of a special enrollment period, but only for a qualified coverage loss, such as the termination of employer-sponsored coverage. States are also taking varying approaches in offering grace periods for non-payment of premiums and in offering special coverage of COVID-19 related services [26].

It is important to consider how individuals who are enrolling in safety net programs are accessing services during a pandemic. Given the closure of many places of business, combined with individuals’ potential hesitancy to seek in-person services based on their own perceived coronavirus risk or caregiving needs, individuals may be challenged to visit social services agencies for enrollment or customer service. While some customer services were expanded by phone and online, there have been widespread challenges enrolling in programs like unemployment insurance [27]. Lower Medicaid enrollment in some states may also reflect the challenges of navigating remote eligibility and enrollment processes since the pandemic.

Our study is subject to several limitations. First, the COVID-19 pandemic is a rapidly evolving situation and data collected in late-April 2020 may not generalize to issues and concerns arising in more recent periods. Since April, the pandemic spread widely across the United States, including to many southern and Midwestern states with relatively weaker safety nets. However, data from April provides an important baseline to assess the evolving need of affected Americans, and to ultimately assess economic recovery. Second, the size of our survey sample (N = 1,007) limits our ability to examine specific subgroups, such as individuals residing in communities that are COVID-19 hotspots. It also limits our statistical power to detect potentially important differences across subgroups. Future research, with larger sample sizes, could beneficially compare differences in program participation based on residence in areas experiencing higher COVID-19 case rates [28]. Third, as with all surveys on public program participation, there is likely to be some under-identification of program participation [29]. Self-reported program participation may be particularly problematic for some programs such as SNAP [30], although we have no reason to believe that reporting bias will be differential between people who experienced COVID-19 employment loss versus those that did not. Fourth, we only asked about past and current use of safety net programs. While asking about future enrollment may be more informative for projecting the burden on the safety net, future intentions are not perfectly predictive of actual enrollment, because of considerable constraints all people face in changing behavior, particularly when considering engaging with complex public programs [31]. Fifth, the data are collected from a nationally representative household panel of English-speaking survey respondents with a modest sample size and thus the study does not offer insights into populations who may be most adversely affected by COVID-19, such as people with limited English proficiency and those experiencing homelessness, in long-term care institutions, and in jails or prisons.

Conclusion

COVID-19 has created historic public health and economic crises. Our survey highlights the hardship experienced by Americans who experienced loss of employment and income early in the pandemic. The crisis has also revealed the cracks in safety net programs, which have been largely underfunded. While some emergency funding has been provided in the wake of COVID-19, the cumbersome and state-specific mechanisms to assess eligibility and facilitate enrollment have created bottlenecks and long wait times across the country. We also identified gaps in the knowledge of existing safety net programs and a need for targeted funding for outreach and enrollment activities. These programs are critical in affording assistance to those most in need to provide for basic needs of food, housing, and income support. A strong economic recovery will crucially depend on how effectively these programs can be streamlined and sustained during this difficult time.

(DOCX)

Click here for additional data file.

29 Jul 2020

PONE-D-20-20058

Access and enrollment in safety net programs in the wake of COVID-19: A national cross-sectional survey

PLOS ONE

Dear Dr. Saloner,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

In particular, please address reviewer concerns regarding presentation of data in the Tables, especially Table 2 which is very difficult to read/interpret.  Regarding concerns about generalizability expressed by Reviewer 2, while the authors certainly cannot control the length of time submissions are under review, perhaps the question of generalizability (or lack thereof) could be framed within a broader context.  Finally, further discussion about challenges accessing safety net services during a pandemic could be more fully fleshed out in the revised manuscript.

Please submit your revised manuscript by August 18, 2020. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Nickolas D. Zaller

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information. Moreover, please include more details on how the questionnaire was pre-tested, and whether it was validated; and clearly report the number of respondents and the response rate.

3. Please correct your reference to "p=0.000" to "p<0.001" or as similarly appropriate, as p values cannot equal zero.

4. In your Methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. Please ensure you have provided sufficient details to replicate the analyses such as: a) the recruitment date range (month and year), b) a description of any inclusion/exclusion criteria that were applied to participant recruitment, c) a table of relevant demographic details, d) a statement as to whether your sample can be considered representative of a larger population, e) a description of how participants were recruited, and f) descriptions of where participants were recruited and where the research took place.

5.We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

6.PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This study reports the results of a national survey conducted in April 2020 to assess the economic conditions of individuals with job loss related to the COVID-19 pandemic. This is a remarkably fast turnaround for such data and the authors are to be commended for their ability to implement these questions in the field on such a short timeline. The results are interesting, but there presentation could be clearer, and there are interesting discussion points that were not address.

Major comments:

1. Table 2 is the potentially most impactful table in the manuscript but it is difficult to interpret. I would recommend three sections in the table: 1) Enrolled before the pandemic; 2) Enrolled after the pandemic; 3) Change in enrollment. As it currently stands, it’s very difficult to understand whether the enrolled before and applied/enrolled since the pandemic are additive.

2. Discussions points: Even before the pandemic, those who experienced job loss were more likely to need help from a safety net program. This speaks to the vulnerability of those impacted by COVID-19 and should be highlighted. Further, only 6.6% of individuals with job loss enrolled in Medicaid. I would expect it to be higher during a public health crisis and may speak to serious challenges in accessing public services during the pandemic.

3. The last sentence of abstract could be interpreted as individuals are “stretching the safety net programs” but this is what the programs were designed to do. Rephrase to indicate that safety net systems will remain critical until the pandemic is over and should be protected from funding cuts. Please correct similar language in the discussion. For example: “Food assistance programs are of particular concern” - these programs not “of concern” but I do agree that their stability is imperative. The problem is not that they are being used - it’s that the US didn’t develop the programs in a way that could respond to large recessions/depressions. A well run program would be able to support individuals during downturns in the economic cycle.

Minor comments:

1. Rather than comparing those with job loss to those without job loss who enrolled, would be more interesting to highlight how much those with job loss contributed to increases in use of safety net services. This could be done with a simple diff-in-diff model to determine how much of the increase in safety services pre and during COVID was associated with job loss.

2. To build on this comment, comparisons between those with and without covid job loss are not wrong but are somewhat odd. Of course those who lost their job would rely on safety net health systems more than individuals who did not lose their jobs. If the authors believe this is the most appropriate comparison in the data, then a clear justification for why such a comparison is important is warranted.

3. Lines 147-152 - The first part of the sentence refers to those with and without job loss. Unclear whether the later part of the sentence related to general population or only those with job loss. Please clarify.

4. There is substantial commentary on Medicaid in the discussion, but would shorten and refocus on what can be gleaned from the study’s findings. For example, what does it mean for budgets, health care delivery, health system budgets, etc that millions more individuals will be relying on Medicaid?

Reviewer #2: I uploaded my review as an attachment rather than pasting it here.

**************************************************************************************************************************************************************

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Charles J. Courtemanche

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: plos one.docx

Click here for additional data file.

16 Aug 2020

Revision Memorandum: Access and enrollment in safety net programs in the wake of COVID-19: A national cross-sectional survey (PONE-D-20-20058)

Editor Comments:

We thank the Editor for the opportunity to revise our manuscript. Below, we respond in detail with bulletpoints to each of the Editor’s comments (shown in bold).

Please address reviewer concerns regarding presentation of data in the Tables, especially Table 2 which is very difficult to read/interpret.

• Based on both Reviewers comments on the former Table 2. We have decided to split the table into two new tables, following a suggestion from Reviewer 2. Table 3 now shows the overall awareness of the programs for the sample, the percent enrolled pre-COVID, the percent enrolled after COVID, and the change. Table 4 displays the differential effect of COVID-19 on enrollment changes between those who did versus did not experience employment reduction. We believe this table is now clearer to readers.

Regarding concerns about generalizability expressed by Reviewer 2, while the authors certainly cannot control the length of time submissions are under review, perhaps the question of generalizability (or lack thereof) could be framed within a broader context.

• We agree this is a valid concern. We have updated our paper to address changes in context that could affect generalizability:

o “Further, data from early in the pandemic provides an important baseline for evaluating the evolving changes in program participation and hardship among vulnerable individuals.” (line 95-97)

o “Since April, the pandemic spread widely across the United States, including to many southern and midwestern states with relatively weaker safety nets. However, data from April provides an important baseline to assess the evolving need of affected Americans, and to ultimately assess economic recovery.” (lines 304-307)

Finally, further discussion about challenges accessing safety net services during a pandemic could be more fully fleshed out in the revised manuscript.

• We agree. Accordingly, we have added text as follows:

o “It is important to consider how individuals who are enrolling in safety net programs are accessing services during a pandemic. Given the closure of many places of business, individuals may be challenged to visit social services agencies for enrollment or customer service. While some customer services were expanded by phone and online, there have been widespread challenges enrolling in programs like unemployment insurance.[27] The lower than expected initial enrollment in programs like Medicaid may also reflect the challenges of navigating eligibility and enrollment processes since the pandemic.” (lines 294-300)

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

• We have consulted these formatting guidelines and we are now consistent with the guidelines.

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information. Moreover, please include more details on how the questionnaire was pre-tested, and whether it was validated; and clearly report the number of respondents and the response rate.

• To provide more transparency for readers, we now include the survey instrument as a supplemental.

• We have clarified that the survey instrument was not pre-tested: “The items that were used in the survey were developed by our team for the purpose of this study; items were not piloted before being used in the study.” (lines 136-137)

• We have provide details on response rate: “The overall response rate for the panel is about 34.0% (American Association for Public Opinion Research [AAPOR] response rate three).[14]” (lines 111-112)

3. Please correct your reference to "p=0.000" to "p<0.001" or as similarly appropriate, as p values cannot equal zero.

• We have replaced these to now read “p<0.001”.

4. In your Methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. Please ensure you have provided sufficient details to replicate the analyses such as: a) the recruitment date range (month and year), b) a description of any inclusion/exclusion criteria that were applied to participant recruitment, c) a table of relevant demographic details, d) a statement as to whether your sample can be considered representative of a larger population, e) a description of how participants were recruited, and f) descriptions of where participants were recruited and where the research took place.

• We have updated our methods section to provide greater detail about participant recruitment and demographics:

o a) the recruitment date range (month and year): “For this study, our team developed the State Health Access Data Assistance Center (SHADAC) COVID-19 Safety Net Survey and contracted with NORC to add the survey questions to the survey that was in the field April 23 to April 27, 2020.” (lines 120-122)

o b) a description of any inclusion/exclusion criteria that were applied to participant recruitment: “We contracted with NORC to administer the survey to a target of 1,000 respondents. The study was restricted to people over age 18. The final sample included 1,007 adults.” (lines 123-124)

o c) a table of relevant demographic details AND d) a statement as to whether your sample can be considered representative of a larger population: “Table 1 shows the demographic characteristics of the study sample. NORC develops weights to national census benchmarks and balances by gender, age, education, race/ethnicity, and region. The weighted sample is similar to a national sample of adults: 51.4% of the sample was female, 44.8% between age 18 and 44, 37.4% non-white, 46.1% with a chronic condition, 36.2% with a high school degree or less, and 83.8% residing in metropolitan areas.” (lines 124-129)

o e) a description of how participants were recruited AND f) descriptions of where participants were recruited and where the research took place: “The AmeriSpeak panel is recruited using stratified, address-based sampling methods that cover approximately 97.0% of all residential addresses. The multi-stage probability sample is created using a national frame area where blocks are sampled from within defined metropolitan or rural areas. AmeriSpeak oversamples in areas with a higher concentration of young adults and minorities and engages in additional efforts to follow up with households that initially do not respond. Individuals are recruited to the panel using a combination of US mail, telephone interviews, and in-person field interviews. Households can respond to the survey by internet (including on smartphones) or by telephone interview. About 85% of the interviews are completed online and 15% are conducted over the phone. The phone option is offered to allow “net-averse” households to participate.” (lines 108-117)

5.We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

• As earlier indicated, we will place the data in repository at acceptance and will provide the DOI to access the data.

6.PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

• ORCID iD for the corresponding author, Brendan Saloner is now provided: https://orcid.org/0000-0001-9013-3023

Reviewer 1

We thank Reviewer 1 for detailed comments. Below, we respond in detail with bulletpoints to each of the Reviewer’s comments (shown in bold).

This study reports the results of a national survey conducted in April 2020 to assess the economic conditions of individuals with job loss related to the COVID-19 pandemic. This is a remarkably fast turnaround for such data and the authors are to be commended for their ability to implement these questions in the field on such a short timeline. The results are interesting, but there presentation could be clearer, and there are interesting discussion points that were not address.

• Thank you for these encouraging words and for the comments, which we believe have strengthened the clarity and depth of the paper.

Major comments:

1. Table 2 is the potentially most impactful table in the manuscript but it is difficult to interpret. I would recommend three sections in the table: 1) Enrolled before the pandemic; 2) Enrolled after the pandemic; 3) Change in enrollment. As it currently stands, it’s very difficult to understand whether the enrolled before and applied/enrolled since the pandemic are additive.

• We agree. Based on your comments and those of Reviewer 2, we have split Table 2 into 2 tables, which are now Tables 3 and 4. Table 3 shows the categories you suggested for the full sample. Table 4 shows the differential changes since the pandemic for the group that experienced employment reduction, similar to the diff-in-diff you suggested below in “minor comments.”

2. Discussions points: Even before the pandemic, those who experienced job loss were more likely to need help from a safety net program. This speaks to the vulnerability of those impacted by COVID-19 and should be highlighted. Further, only 6.6% of individuals with job loss enrolled in Medicaid. I would expect it to be higher during a public health crisis and may speak to serious challenges in accessing public services during the pandemic.

• In the Discussion, we now highlight the point that those who experienced job loss were more likely to need help from a safety net program prior to the pandemic: “Overall, our study findings highlight the precarity of individuals who were most affected early in the COVID-19 pandemic” (lines 213-214)

• We also more explicitly address the issue of low initial uptake of Medicaid early in the pandemic and how this might reflect challenges accessing public services: “Lower than expected initial enrollment in programs like Medicaid may also reflect the challenges of navigating eligibility and enrollment processes since the pandemic.” (lines 296-298)

3. The last sentence of abstract could be interpreted as individuals are “stretching the safety net programs” but this is what the programs were designed to do. Rephrase to indicate that safety net systems will remain critical until the pandemic is over and should be protected from funding cuts. Please correct similar language in the discussion. For example: “Food assistance programs are of particular concern” - these programs not “of concern” but I do agree that their stability is imperative. The problem is not that they are being used - it’s that the US didn’t develop the programs in a way that could respond to large recessions/depressions. A well run program would be able to support individuals during downturns in the economic cycle.

• We strongly agree and regret that our prior phrasing did not emphasize the importance of stabilizing safety net programs to meet the needs of vulnerable people. We have rephrased in several places.

o For example, last line of abstract: “The economic devastation from COVID-19 increases the importance of a robust safety net.” (lines 50-52)

o In the Discussion: “The stability of food assistance programs is of particular concern.” (line 237)

o “The affordability of housing is another concern.” (line 250)

Minor comments:

1. Rather than comparing those with job loss to those without job loss who enrolled, would be more interesting to highlight how much those with job loss contributed to increases in use of safety net services. This could be done with a simple diff-in-diff model to determine how much of the increase in safety services pre and during COVID was associated with job loss.

• Consistent with your proposed modification to Table 2, we have created Table 4, which has the layout that captures this “difference-in-differences” style effect.

2. To build on this comment, comparisons between those with and without covid job loss are not wrong but are somewhat odd. Of course those who lost their job would rely on safety net health systems more than individuals who did not lose their jobs. If the authors believe this is the most appropriate comparison in the data, then a clear justification for why such a comparison is important is warranted.

• Thanks for pushing us to clarify this. Our point is consistent with your suggestion #1: to demonstrate the disproportionate need among this group. We state this more clearly in the manuscript:

o “Because we were interested in identifying the disproportionate changes in program participation among those experiencing employment reduction, we fit a regression model for program participation that estimates the average change since the pandemic, the baseline rate for those without employment reduction, and an interaction term. This interaction term is analogous to a difference-in-differences coefficient, representing the change in program participation since the pandemic for those with employment reduction versus those who without employment reduction.” (lines 150-156)

3. Lines 147-152 - The first part of the sentence refers to those with and without job loss. Unclear whether the later part of the sentence related to general population or only those with job loss. Please clarify.

• We have clarified this section of the paper.

4. There is substantial commentary on Medicaid in the discussion, but would shorten and refocus on what can be gleaned from the study’s findings. For example, what does it mean for budgets, health care delivery, health system budgets, etc that millions more individuals will be relying on Medicaid?

• We have reduced the discussion related to the specific provisions of the Medicaid program and have refocused on the study findings as follows:

o “The increased enrollment in public programs could have important implications for state budgets and delivery system. Early in the pandemic, state Medicaid programs began projecting that the weakened economy would increase program enrollment and total spending.[23] Testing and treatment related to COVID-19 could also contribute to rising Medicaid spending, though this spending could be offset by other forms of medical care that decreased after the pandemic, such as preventive office visits. State programs have a variety of policy options for easing transitions of new members into Medicaid, reducing churn, and simplifying enrollment. A recent analysis suggests that Medicaid enrollment increased early in the pandemic, although these increases were not correlated with enrollment changes in those states [24]” (lines 275-283)

Reviewer 2

We thank Reviewer 2 for detailed comments. Below, we respond in detail with bulletpoints to each of the Reviewer’s comments (shown in bold).

This paper is a straightforward comparison of economic hardship and enrollment in safety net programs between those who experienced COVID-19-related job loss (defined as either complete job loss or a reduction in hours) versus those who didn’t. Data come from a special COVID-19 economic impact questionnaire that the authors had appended to the April 24-26 survey of the AmeriSpeak panel. The key results are: 1) 28% of the sample experienced job loss, 2) there were sizeable differences in program participation even before the pandemic between those who would eventually experience job loss and those who would not (i.e. job loss hit the lower end of the labor market harder), and 3) those differences became even larger after the job loss happened, 4) there were huge differences in subsequent priorities for stimulus check spending and ability to pay for basic needs. There is also an attempt to examine the influence of demographic characteristics on these results, though the estimates are generally too imprecise to be useful.

• Thank you for this helpful synopsis of our paper which we agree touches on key points of interest.

These results are unsurprising and the analyses are fairly simplistic (e.g. no attempt to leverage exogenous variation to identify the causal effects of job loss on the various outcomes, rather than just its associations). That said, the data source is novel and there is certainly value to quantifying the magnitudes of the associations. The more attention that can be brought to the need for a robust social safety net and continued stimulus during this time of crisis, the better.

• Thank you, we agree that the paper is entirely descriptive and that we do not exploit exogenous variation. Per your comment, our main intention in this paper is to demonstrate the magnitudes of need for services early in the pandemic. We have added a suggestion in the paper for further work to examine exposure to COVID-19 cases:

o “Future research, with larger sample sizes, could beneficially compare differences in program participation based on residence in areas experiencing higher COVID-19 case rates.[28]” (lines 308-310)

Comments/suggestions:

1) I don’t see the value of the regressions in Table 5. As the authors themselves concede, the confidence intervals tend to be too wide for the results to be terribly informative. I would just drop this table, freeing up space for better exposition in the text and potentially also allowing the clunky Table 2 to be split into multiple exhibits (more on that next).

• We agree that these results are not that informative because of the imprecision and have removed the regression table.

2) It took me a while to figure out exactly what’s going on in Table 2. The confusing part is that there’s really two very distinct questions being asked:

a) How different was baseline enrollment (i.e. pre-job-loss) between those who would eventually lose their jobs/have hours cut versus those who would not? If I understand correctly, the value of those results is in showing that the people who got laid off/had their hours cut were disproportionately “vulnerable” even prior to the pandemic.

b) How different was new enrollment during the pandemic between those who experienced job loss/reduction versus those who did not? This is more like the usual difference-in-difference-style question.

Since these are such distinct questions, I think it would be less confusing if they were treated more distinctly throughout the paper. For instance, split Table 2 into two separate tables, and better explain their distinctiveness in the abstract, intro, and body of the paper. To provide one example of how the distinction is currently murky, the abstract says, “Those who experienced COVID-19 job loss versus those who did not were significantly more likely to have applied or enrolled in >1 program … and also significantly more likely to specifically have enrolled in unemployment insurance … and SNAP.” More likely to have enrolled when? Nothing in that sentence specifies new enrollment during the pandemic. So it ends up reading as a continuation of the previous sentence about baseline differences.

• Thank you for these helpful interpretative points, which dovetail with concerns expressed by Reviewer 1. We regret that the table was confusing in the prior draft. Based on your suggestions, we have split the former Table 2 into two pieces: Table 3 that shows the overall rates of program participation before the pandemic, since the pandemic, and the change and Table 4, which represents the difference-in-differences, and is more about the contrast between those who experienced and did not experience employment reduction.

3) One substantive critique is that the authors ignore the well-known issue of misreporting in survey-based program participation measures. While the critique applies to all programs, the literature on misreporting in SNAP is especially well-developed. A recent paper that provides a detailed discussion of that literature is https://onlinelibrary.wiley.com/doi/abs/10.1002/soej.12364

• Thank you for raising this very reasonable issue. This reporting bias is likely to lead us to understate program participation overall, but we have no a priori reason to believe that such bias is likely to be greater among people with versus without COVID-19 related job loss. We added further context on this and included the helpful suggested citation:

o “Self-reported program participation may be particularly problematic for some programs such as SNAP,[30] although we have no reason to believe that reporting bias will be differential between people who experienced COVID-19 employment loss versus those that did not.” (lines 311-314)

4) Given the rapidly changing nature of the pandemic, a survey from late April is already a bit dated. The authors acknowledge that “data collected in late April 2020 may not generalize to issues and concerns arising in more recent periods.” I appreciate the caveat but they should push harder here. If the results don’t generalize, then why do we care? At lease some argument needs to be made for generalizability.

• Very fair point. We expand on this issue in the paper, arguing that the early experience in the pandemic can offer an important baseline for further study. The areas that were hardest hit during this time were northeast states with stronger safety net programs, whereas the pandemic in summer 2020 evolved to more directly affect areas with weaker safety net programs. We have added additional text on this issue:

o “Further, data from early in the pandemic provides an important baseline for evaluating the evolving changes in program participation and hardship among vulnerable individuals.” (line 95-97)

o “Since April, the pandemic spread widely across the United States, including to many southern and midwestern states with relatively weaker safety nets. However, data from April provides an important baseline to assess the evolving need of affected Americans, and to ultimately assess economic recovery.” (lines 304-307)

5) p. 13: “There was no statistically significant difference in participation in at least one safety net program between people who subsequently experienced COVID-19 job loss than those not experiencing job loss (50.0% versus 37.7%, p=0.120).” I don’t like this sentence. The magnitude of that difference is large, and the p-value is close to the threshold … more nuance is needed.

• Yes, we agree that the prior statement was overly focused on the p-value. However, we are no longer making this comparison in the paper, so this has been removed.

6) I’d give some thought to whether “job loss” is the best term to describe the “treatment” group. It’s actually the combination of those who lost jobs completely and those who had their hours reduced – and the majority is the latter. It’s really more of an “adverse employment shock” group. Calling it “job loss” may be catchier but it is a recipe for being misquoted in the media.

• This is an astute observation. We have renamed this group “employment reduction” and more clearly indicated what we mean by this term.

21 Sep 2020

Access and enrollment in safety net programs in the wake of COVID-19: A national cross-sectional survey

PONE-D-20-20058R1

Dear Dr. Saloner,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Nickolas D. Zaller

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have satisfactorily addressed my concerns. These data will be a nice contribution to our understanding of the economic and social impact of COVID-19.

One very minor comment - Job loss should be changed to employment reduction in Tables 5 and 6.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Charles J. Courtemanche

28 Sep 2020

PONE-D-20-20058R1

Access and enrollment in safety net programs in the wake of COVID-19: A national cross-sectional survey

Dear Dr. Saloner:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Nickolas D. Zaller

Academic Editor

PLOS ONE