This page contains a calculator which can be used to design a guaranteed basic income or a negative income tax in the United States, using hard Census data and spending data from the Congressional Budget Office. In its basic form, the tool only requires two inputs to calculate the costs of various forms of a guaranteed basic income (or “negative income tax”):
- A guaranteed basic income (in dollars) for every American over the age of 18 years old.
- An effective tax rate to limit those benefits as income is earned.
(I explain the difference between the negative income tax and the guaranteed basic income in the FAQ section).
In its advanced form, the tool allows you to pick and choose from line items in the CBO’s 2016 Fiscal Outlook. It automatically tallies that spending and compares your proposed negative income tax to the cost of the currently active programs in the United States.
Be sure to share your transfer program in the comments section, and to tell people about this article and calculator!
III. Notes & Disclosures
V. Frequently Asked Questions
VI. Citations, Code Use and Data
Introduction to the Negative Income Tax and the Guaranteed Basic Income
Why create this calculator?
Simple: embarrassment at the quality of discourse over the Guaranteed Basic Income in the press. We’ve seen altogether way too many pieces which simply quote a number (say: $10,000?) and multiply that number by the adult population of the United States (242,248,036 +/- a few) to come up with a “cost” of a negative income tax.
A properly designed guaranteed basic income or negative income tax has much more nuance than penciling in $2,422,480,360,000.00 to a story and waiting for clicks.
Witness the United States’ most successful welfare program in its history, the Earned Income Tax Credit. The American EITC can trace its roots to the philosophy of Milton Friedman. Friedman noted that if we must have taxes, we should still design the tax code in a way to minimize harm to worker incentives. His solution was simple: subsidies to the poorest, with a phase out in benefits as a person made more income.
Since the CBO is now looking at an almost identical problem (the disincentive to work related to the drop-off in subsidies and transfers in the Patient Protection and Affordable Care Act), the time has come to look at cleaning the atheroscleritic transfer and welfare programs in this country with a sleek, easily administered negative income tax.
For more on the Negative Income Tax (especially as it relates to the Minimum Wage), please see our previous articles on the subject:
- Comparing Forms of Entitlement Programs (and Part II)
- The Minimum Wage Mistake
- A $9 Minimum Wage is the Wrong Policy
- The Minimum Wage Crushes the Teen Employment Rate
- The Moral Case Against Price Controls
For more calculators and original studies, please visit our Calculations and Visualizations page for an listing of what we’ve produced.
Results: How Much Would a Guaranteed Basic Income Be?
We had to populate the calculator with something, so taking all of the caveats of the next section into account, we targeted a 50% phaseout/effective tax rate and hitting the current system’s cost within 1 or 2%. We feel there would be a lot of administrative cost savings on top of just the program elimination, but since our $14,400 number will be quoted most often we’ll leave it in place and just tell you that “it’s probably higher”.
But remember, it’s a calculator. Not only can you roll your own program, you can also adjust the programs you’d like to eliminate with your program. Insert your own opinions and biases and hit ‘Calculate’, as often as you’d like!
Our base case is a good one, anyway. A $14,400 guaranteed basic income to Americans comes in at right above $2.1 Trillion in costs, and is roughly balanced by eliminating all the programs we’ve selected by default in the ‘Advanced Options’ section..
The Notes: Decisions, Assumptions, and Caveats
For the 2015 update, of the 300+ million or so Americans, we chose to apply our analysis to the roughly 242,248,037 adults in the country at that time. The current EITC has modifiers for the number of children at home, so any adjustments for children would come at an additional cost, or would have to be deducted from the single families. Of course, that may not matter; with our default subsidy of $14,400, the average non-working parent would have more resources than in the current regime.
We decided for our default cost calculation, we would keep government employee benefits in place. We recognize that this particular issue will be a hot-button one, so you can adjust it in “Advanced Options” to fit your opinion of whether a guaranteed basic income should replace those programs. (Our take? Those benefits were a condition of employment for the government, and likely an important part of the employment decision on the part of the worker in question.)
It’s important to note that the analysis we did was a static analysis. If you’re unfamiliar with the term, it means that we assumed no changes in behavior between the tax code and transfer system in place as of today and a proposed future system. On balance, we believe that a negative income tax and guaranteed basic income would be a net positive for putting in place incentives to work:
- Increased risk taking – a safety net of $14,400 (or $28,800 for a 2 adult household)
- Reduced confusion – you need only calculate your income to know what benefit is available
- Reduced fiscal disincentives – easier to avoid program conflict and 100%+ effective rates (witness this chart of Pennsylvania’s disjointed programs)
- Easier Enforcement – no maze of programs to verify eligibility
So, in our opinion, this is a worst case analysis. If you think that the current system with dozens and dozens of repeated programs with no clear jurisdiction, fiscal traps, eligibility problems, abuse, fraud, and myriad other problems is better for encouraging work (and you think the Pennsylvania picture linked above isn’t a big deal)? Well, the data is at the bottom and we’re releasing it with the code as open source – feel free to build on what we started and toss us a link when you’re done.
Another important note – the CBO’s cost summary buries this footnote: “Data on spending for benefit programs in this table generally exclude administrative costs, which are discretionary”. That means that if you think this will be easier to administer than the current regime, you’ll be able to push the benefits past $14,400 per adult.
We’d say that’s a fair bet (see our list in this section). CBO, care to weigh in on what the full cost would be, with administrative costs clawed back?
Methodology on the Negative Income Tax and Guaranteed Basic Income Calculator
Excuse our verbosity, but we’d like to make this easy for as many people as possible to reproduce. Consider yourself our peer reviewer, and here’s how you replicate this work:
- As so many of our projects do, this one started with a trip to IPUMS-CPS. It’s a judgement call, but we considered current transfer payments to be Social Security Income, SSI Income, Unemployment Benefit Income, and Welfare Income.
- Open your favorite stats program, and eliminate everyone with an age under 18.
- Subtract the four categories from total income to form an ‘earned income’ type column.
- Find a way to limit your points – we used the wtd.quantile program in R, in the package Hmisc. Another good option would be random sampling (be sure to recompute your weights if you do this!).
- Why did we cull? Because we need to find a best fit function. I use ZunZun, which is better than you eyeballing it and pretending you know which type of function best fits the data (Editor: unfortunately, the developer had to remove ZunZun for health reasons – you can run your own version of the site though– that’s what we did for this version). You need to do it in both directions – from income to index, and from index to income. (At this point, I should mention my fit is good for around 97.5% of incomes, but again, is sloppy on the top end – I’m willing to assume we won’t give subsidies to the top 2.5%, so this is an acceptable tradeoff in my opinion).
- Now, do calculus – you need to find the integral of: (BASIC INCOME) – (EFFECTIVE TAX)*(your “index to income” function)
You’re done, but here’s how to use it:
- The maximum income affected is found by “Basic Income” / “Effective Tax” (our ‘phase out’ percentage).
- Find the index of the maximum income from step one using your “income to index” function.
- Integrate from ‘0’ to ‘Highest Index’ over the function you found in step 6 above.
- If you used 0 – 242,248,037 as your indices, you’re done – that’s the cost. If you used a different number, apply weights so it’ll equal 242,248,037 to get your cost.
If a part doesn’t make sense, you can review the guaranteed basic income calculator code – it has a fair number of comments.
Frequently Asked Questions
Since I wrote this section with the article, I’m just anticipating questions. If you ask something, I reserve the right to add your question to this section so the answer will benefit future tax-code designers.
- “What’s the difference between a Guaranteed Basic Income and a Negative Income Tax?”
In theory, there is absolutely zero difference between the two. In practice, they differ only in how the benefits would be distributed:Guaranteed Basic Income: Everyone is given a check for a certain amount, and the United States will claw back money from higher earners using taxes. The effective benefit is exactly the same except people might have some loss aversion since they have to give money back.Negative Income Tax: The guaranteed income is treated as a credit, so other income is computed first and deducted from the check before it is given to adults. The effective benefit is exactly the same except it might be psychologically better as people never ‘had’ the full benefit amount, needing to send it back in taxes.
- “I’m new here – what other material would I like if I like this?” We remember fondly replacing the Federal Reserve with a robot which, ironically, pulled data from… the Federal Reserve Bank of St. Louis. Go redo our economic policy with our automated Taylor Rule Calculator after you’re done with our welfare system. Then nap – that’s a lot of redesigning for one day.
- “Why did you use individuals instead of households?” Simple: incentives. Just like when the IRS mandated Social Security numbers for dependents and millions of phantom children disappeared overnight, basing these benefits on households would lead instantly to millions of newly formed households. We can differ on how to treat children, but basing this program on households will make our assortative mating issues worse for those who make less than our threshold. Since I know this point will make some angry, please leave your Rube Goldberg method of reducing this incentive in the comments.
- “What about cost of living?” What about it? Remember: most areas of the country have 3 or 4 layers of governance (at a minimum – ask someone with an HOA for more details). There is nothing stopping, say, New York City from designing their own Negative Income Tax. This data applies to the federal level… let’s not ask people in Kansas to subsidize us folks living in Silicon Valley, okay?
- “But what if we don’t like how people spend their money?” This one gets asked a lot when it comes to articles like this. Let me ask: how is it humane to attach strings to welfare dollars and insist that an omniscient program designer is better able to decide how someone should spend their money than that someone? This doesn’t assume a rollback of laws – child neglect would still be a crime, for example. The point is the establish a basic income at a level which is a bit uncomfortable – people will be motivated enough to try to better themselves through employment, since you won’t hit the benefit cliffs of the current system (see: Casey Mulligan’s work for more on that subject). Benefits are unreliable anyway – do you really think people don’t trade food stamps for other goods? If yes, here’s a link to Reddit’s economics subreddit… you’ll fit right in.
- “Where are the 0% and 100% categories in the graph?” They skew the graph too much – 0% is negative and 100% is over $1,000,000 (We know that some Americans make well over that). Rest assured: the equation takes them into account.
- “What Equation Did You Use To Find the Income of an American Adult at a Percentile?”
(It’s very good up to around 97.5% – so we cap the tool there.)
|y = a + bx^1 + cx^2 + dx^3 + fx^4 + gx^5 + hx^6 + ix^7 + jx^8|
|a = -764.99982159273|
|b = 225.935125029779|
|c = -0.615201717018447|
|d = -3.0554342274265|
|f = 0.231633424887472|
|g = -0.00665845587364395|
|h = 0.0000943329859424975|
|i = -0.000000664300893618064|
|j = 0.00000000187218267553746|
- “But How Do You Get the Total Income Up to that Percentile for Your Estimates?”
Do calculus, of course!Just integrate the above equation or ask Wolfram Alpha or someone to do it for you (it’s a polynomial though, you should be able to do it or learn, simply!). Then take the integral from 0 to your number. (It’s already expanded in the tool’s source code – search for the function ‘incomeIntegrate’ to see it).
- “Are you sure…. well, did you consider…?” Click here. It’s right.
- “Why did you make this?” I thought I answered that above, but call it intellectual curiosity, and a shot at one of you hiring me to build a better one. Or sending me some paid-for grad students to do the research part for me so I can concentrate on the calculator – either is a great result.
- “How long did it take?” IN 2013: Roughly 10-12 hours – and it would have been 20 if I didn’t write a lot of the helper functions in my other calculators. The breakdown was ~8 hours on the research, 2 on the calculator styling, and 2 on this explanation article. If I did it again, I could do the research in an hour or two (there is a learning curve to the process, trust me).
IN 2015: The same amount of time, except 90% was dedicated to getting ZunZun’s source code running on my MacBook. Here’s a hint: qsize() for queues gives a NotImplementedError on the Mac. (I might clean up my code and post it on GitHub).
Since the calculator was already made, that part took like 1 hour this time.
- “What the [heck] is a data scientist?” I’ve been told it’s a programmer who can’t code mixed with a mathematician who can’t do math. I don’t claim the title, anyway: I’m a computer engineer, pure and simple.
Citations & Data and Code License
Find the 2015 annual income percentile data here. Just hit ‘view source’ on the calculator for that part of the code.
Using and quoting us doesn’t change for this piece – you are free to quote excerpts of this article, or to quote the calculator (or produce screen shots of the calculator) with a link back to Don’t Quit Your Day Job, whether you have compliments, comments, questions or criticism. The contract only applies on the code/software front:
You are free to use, analyze, and modify and create derivative works of the code and/or data, as long as you cite Don’t Quit Your Day Job… as your source.
Additionally, you agree the code and data is provided AS IS, with no implied warranties expressed or implied. In no event shall we be liable for any claim, damages, or other liability whether in an action of contract, tort, or otherwise, arising from, out of or in connection to our software or data.
- IPUMS-CPS: Miriam King, Steven Ruggles, J. Trent Alexander, Sarah Flood, Katie Genadek, Matthew B. Schroeder, Brandon Trampe, and Rebecca Vick. Integrated Public Use Microdata Series, Current Population Survey: Version 3.0. [Machine-readable database]. Minneapolis: University of Minnesota, 2010.
- Congressional Budget Office
- ZunZun source code (run it on a machine you own)
- R (Stats Program)
“[A Guaranteed Basic Income] would do more efficiently and humanely what our current welfare system does so inefficiently and inhumanely. It would provide an assured minimum to all persons in need regardless of the reasons for their need while doing as little harm as possible to their character, their independence, or their incentive to better their own condition.”
– Milton Friedman, Free to Choose
It’s time for all the vested interests to recognize that we can design a system which both eliminates the moral hazards, ridiculous inefficiencies and unintended consequences of our current system, while showing compassion to those who have need regardless of cause.
A guaranteed basic income, designed properly to not discourage work, is exactly that system. Here’s the proof you wanted. Use it in good health, and please share with us the programs you’ve eliminated and the benefits levels you’ve tested. Thanks for reading!
Comments? Questions? Suggestions? Leave them below.