How many Americans live in a family that receives at least $1,000/adult in monthly govt. benefits?

Factual Questions
1

Good afternoon,

In “Andrew Yang’s UBI proposal” over on the Great Debates forum, Ryan_Liam, Heffalump and Roo and I disagreed over how many Americans would actually benefit from a $1,000 monthly payment minus existing benefits from “welfare programs, food stamps, disability and the like”. I am inclined to think a significant proportion of Americans live in families that already receive at least $1,000 per adult in monthly benefits; my counterparts do not.

The question then is: “How many Americans currently live in a family that receives at least $1,000 per adult in monthly benefits from welfare programs, food stamps, disability and the like?”

The thread has largely winded down and I’m more interested in how to procure a statistic like this than the number itself. I have come to the conclusion that the statistic doesn’t exist, but is there a way to make an educated guess?

~Max

2

First things first, a Google search for the question as written mostly returned results on SNAP requirements. A search for “How many Americans receive over $1,000 in welfare?” wasn’t too helpful either. One site cites the Census Bureau saying the average monthly payment benefit for welfare recipients is $404[1]. A Forbes article in 2015 propogates a rough claim of $25 per day, which is $9125 per year or just over $760 per month[2]. The distinction is explained because the Forbes writer was British and used welfare in a general sense. Apparently here in America welfare means cash assistance, a connotation I was not previously aware of despite being American. I spent quite some time failing in my Google searches - lots of results about universal basic income, too. Though I did not find my answer, most citations lead back to the Census Bureau.

This makes sense because the Census Bureau, when not actually conducting the decennial census, is in the business of doing surveys and producing useful statistics. So I headed over to census.gov and searched there.

A 2015 Census Bureau press release says 21.3% of the population participates in government means-tested assistance programs[3]. The underlying report breaks down average participation by program and by select demographics within each program[4]. It also includes average monthly benefits by program.

I tried combining average participation and benefits and multiple programs to reach over $1,000 but the margin of error was too high to retain meaning. Also average monthly benefits for Medicaid and housing assistance are missing from the report, presumably because the wide variation would render any such statistic meaningless. Now what? Back to Google searches…

Next I found a 2018 report from the Bureau of Labor Statistics[5]. When I first found this I figured I had enough between it and the Census Bureau report to answer my question. I was wrong. In this report also, the benefit amounts for Medicaid and housing assistance are missing. I tried adding figures in Table 2, and it seems that in 2014 0.9794% of 3.037 million single-parent families received, on average, $1,143 in public assistance, SNAP, and SSI. This would be at least 6.074 million people, assuming each single-parent family has one adult and at least one child. But this isn’t accurate at all: I’m missing the margins of error; a family on one program does not have the same chance to be on a second program as any other family. For example, signing up for TANF in my state will also sign you up for SNAP. Neither does it fully answer the question.

I looked at the sources for Table 2 to try and find the margins of error. It turns out the source is not the Consumer Expenditure Survey as implied on the table, it is the very same 2015 Census Bureau report. That explains why Medicaid and housing assistance are missing. So we are back to one source data set with no published statistic that answers my question.

~Max

[1] 45 Important Welfare Statistics for 2019. (2018). LexingtonLaw. Retrieved from https://www.lexingtonlaw.com/blog/finance/welfare-statistics.html
[2] Worstall, T. (May 4, 2015). The Average US Welfare Payment Puts You In The Top 20% Of All Income Earners. Forbes. Retrieved from The Average US Welfare Payment Puts You In The Top 20% Of All Income Earners
[3] 21.3 Percent of U.S. Population Participates in Government Assistance Programs Each Month. (2015). United States Census Bureau. Retrieved from https://www.census.gov/newsroom/press-releases/2015/cb15-97.html
[4] Irving S. K., & Loveless T. A. Dynamics of Economic Well-Being: Participation in Government Programs, 2009–2012: Who Gets Assistance? (2015). United States Census Bureau. Retrieved from https://www.census.gov/content/dam/Census/library/publications/2015/demo/p70-141.pdf
[5] Program participation and spending patterns of families receiving government means-tested assistance. (2018). Bureau of Labor Statistics. Retrieved from Program participation and spending patterns of families receiving government means-tested assistance : Monthly Labor Review: U.S. Bureau of Labor Statistics

3

Next I tried to look at the source data: the Census Bureau sourced their data from their Survey of Income and Program Participation’s 2008 panel. This survey followed some forty-two thousand households over four years[1]. The data is broken up into sixteen waves, one for every four month period. This means interviewers asked each household to complete a survey every four months. Each wave is broken into four rotation groups, which just means interviewers had time to survey about ten thousand households each month and split it up that way.

The data is available for access as either: raw bytes for each wave and a technical manual (FTP), an outdated java applet interface (DataFerrett), or an ambitious research program that anonymizes any confidential information before showing it (Synthetic SIPP)[2].

Each wave core file (there are 16) comes with a thousand-plus-page technical reference[3]. Following that to write a computer program is more work than I am willing to do.

Access to Synthetic SIPP requires a formal application and, due to the experimental nature of the program, the Census Bureau wants to validate any conclusions themselves before publishing.

This leaves DataFerrett, which only works on Internet Explorer and ancient versions of Java. I spent some time using the program, but could not for the life of me figure out how they got the median amount of benefits from that dataset. I am guessing they hide the amount of benefits from public access via DataFerrett. Apparently they asked each household “how much did you receive in government benefits in the past three months” but this was only asked once, for wave 0. I can’t see what they responded, but I can have DataFerrett show me, based on the data that I can’t see, how many families answered with more than a certain amount. Unfortunately I couldn’t find a way to relate that answer with the number of people in the household, and it is unclear whether SIPP tried to verify what was reported.

And now I am at the end of my rope.

~Max

[1] SIPP Users’ Guide, Chapter 2. SIPP Sample Design and Interview Procedures. (2015). United States Census Bureau. Retrieved from https://www.census.gov/content/dam/Census/programs-surveys/sipp/methodology/SIPP_USERS_Guide_Chapter2_2009.pdf
[2] Survey of Income and Program Participation Datasets. (n.d.). United States Census Bureau. Retrieved from Survey of Income and Program Participation Datasets
[3] 2008 Panel Complete Technical Documentation. (2018). United States Census Bureau. Retrieved from https://www.census.gov/programs-surveys/sipp/tech-documentation/complete-technical-documentation/complete-documents-2008.html

4

I suspect you’re right that it doesn’t exist in some easy format, but we could take a stab at it. My contribution:

In 2016 the National School Lunch Program covered all or part of the cost of school lunches (and breakfasts and milk?) for 30.4 million children. source (PDF)

According to the Census Bureau, there were about 62 million “children” (ages 5-19) in 2016 (source)

That makes it roughly half of school-aged children. The amount per adult is going to depend a lot on how many children there are per family (and how many parents), and whether they’re getting breakfast, lunch, milk, or some combination of them.

According to this, most of the school lunches are around $3-$3.50 (getting as high as $5.68 in Alaska!), breakfasts are ~$2 each, and milk is $0.45-$1.48. So, some back of the envelope math, that might be $5-6 per kid per school day, with about 180 school days per year, that alone is right around the $1,000 mark per kid.
BTW, thanks for your contributions in the UBI thread. They were interesting reads.

5

Max S., are you including:

Social Security Retirement Benefits,
Veterans Disability Payments,
Medicare Payments,
and/or Workman’s Comp?

6

I’m impressed at the lengths you’ve gone to try and get this data. Should we all be surprised that this number isn’t a lot more known and transparent? Do politicians have access to any better data to make their determinations?

Also, if you are talking about school lunches why not talk about the cost of school itself? Looks like that costs about $1000 per kid per month itself, or at least close to that ballpark.

7

When CBO looks at income “after taxes and transfers”, I’m wondering where they’re getting their “and transfers” numbers.

8

I’ll add that the assisted living place my wife works at plus where her mother lives and where my father lives the numbers are approaching 100% for the residents that receive any state or federal financial assistance in excess of $1000/mo.
I’d WAG that more elderly receive SS and housing and medical assistance in excess of $1000/mo than their younger poor counterparts.

9

well I get 1075 from ssa via dads retirement which they switched me to from the ssi I was getting but my aunt and disabled cousin get ssa/ benefits because of my uncle passing away and my cousin gets 145 from ssi and like 875 from SSa

California at one time would split your check so youd get the federal half as ssi and the state half as ssa that way when ssi was cut they could point out "the states not cutting your money Washington is "

10

Because the federal government isn’t generally subsidizing the cost of school itself, are they? It’d be outside the scope of the OP’s question about government benefits.

11

They do somewhat. More or less depending on the district. But I’m not sure the range of percentages.

12

Well, I don’t.

There is an opportunity for me to receive some assistance, but I doubt it would amount to a thousand a month, and it would certainly would not be ongoing. But I don’t need it, and am unwilling to jump through bureaucratic hoops for it. Didn’t take the free help on filing my income tax, either.

So, put a no in my column on your survey.

I also won’t claim deductions for charitable donations. They’re charitable. You know, left hand, right hand, mind your own business. Read Matthew. Chapter six I think.

Tris

I earn more than I spend. How rich does a man need to be?

13

USA Facts is a great non-partisan web site that basically aggregates available US government databases. It was founded and funded by former MSFT CEO steve ballmer.

What you’re looking for might be found in the site, although my 2 minute quick searches didn’t pull it up.

The OP and others might find this a great source of interesting data so throwing this out there

14

Just to chime in, but a strict reading of the OP makes me believe the OP is asking how many Americans live in a family where EVERY adult in the family receives $1K/month (>= $12K/year) in “welfare” type payments including disability. I do not know the answer, but if I am reading the OP correctly I suspect the number is VERY small. And of course minors, students in public schools, don’t count by definition of adult. If the OP is asking how many Americans live in a family where one or more adult in the family receives $1K/mnth or more, that is a far larger number and probably what the OP is asking.
Also, would the OP include people on medicaid? If that is to be included, then the numbers will fluctuate wildly month to month as people either need medical care or don’t during that month.

15

Given the thread this one is based on (linked in the OP), you aren’t reading it right. The point is that someone has proposed UBI of $1000/adult. How many households could exchange all their benefits and come out worse-off?

16

OK. I didn’t read the original thread. So my ‘strict’ reading is in fact the correct one.
I have no actual information to contribute but I would be surprised if there are large numbers of households where every member received >= $1K/mnth.

17

I’m surprised no mention of the Earned Income Tax Credit (EITC). From Wikipedia:

“At a cost of $56 billion in 2013, the EITC is the third-largest social welfare program in the United States after Medicaid ($275 billion federal and $127 billion state expenditures) and food stamps ($78 billion). Almost 27 million American households received more than $56 billion in payments through the EITC in 2010. These EITC dollars had a significant impact on the lives and communities of the nation’s lowest-paid working people largely repaying any payroll taxes they may have paid. The Census Bureau, using an alternative calculation of poverty, found that EITC lifted 5.4 million above the poverty line in 2010.”

18

Not every member, every adult member.

About 53% of households with unmarried parents are single mothers raising children on their own, and about 12% of those kinds of households are single fathers raising children on their own. Cite. About 23% of households with children are single mothers with no other adult. Cite. Since

(cite) if you count the benefits for those households as going only to the adult, all told I would expect to find a significant number of such households. Total benefits to each poor household - not per person, per household, and including Medicaid and various other non-cash benefits - is something like $14,000 a year (cite.

It depends on how you slice the numbers, and how UBI is defined, but it does not seem out of the question that there are already a number of adults who are receiving more than $1000 a month, overall and on average. Whether or not UBI is better or worse than that is a question beyond the purpose of GQ.

Regards,
Shodan

19

Not every member. A total of $1k per adult, even if one currently receives $2k and the second receives nothing.

20

In the original thread I posted this which linked to this treasury document. It give some of the information divided by subgroups of:
Non-elderly non-disabled households with children;
Non-elderly disabled households with children;
Non-elderly disabled households without children;
Non-elderly non-disabled households without children; and
Elderly households.

The analysis is not all benefits though.They define it

So a slight underestimate of the benefits per household and not a direct answer to the question in this op (for example it does not specify how many adults are in each household in each decile) but an oblique way of getting some sense of it.