On September 2, the Congressional Budget Office published an estimate of the amount of additional tax revenues the Internal Revenue Service would reap if its budget were increased by a total of $80 billion over a 10-year period.
The answer? $200 billion.
I was stunned by how low that number is.
Think about what that means for the deadweight loss. We normally estimate DWL by looking at how the tax distorts the economy, causing a mix of producer surplus loss and consumer surplus loss. A standard number for the DWL from many taxes is 30 percent of the revenue raised. But the CBO analysis doesn’t even touch this; it’s just about auditing and collection costs.
So if the standard 30 percent applies, this amounts to an overall DWL of 70 percent of the amount of revenue raised. That’s one inefficient way to raise tax revenue.
Is there any factor that could make the 30 percent number be lower? Yes. To the extent people anticipate more scrutiny by the IRS, some might actually reduce their avoidance. They might, for example, declare income that they would have otherwise hid. It takes real resources to do that kind of avoidance, which is really evasion, the illegal part of avoidance. It takes good lawyers and lots of time, for example, to set up a bank account in Grand Cayman. So some people might actually engage in less avoidance and evasion.
Interestingly, though, Phill Swagel, the director of the Congressional Budget Office, goes the other way. He writes:
Second, CBO expects taxpayers to adapt to the IRS’s enforcement activities and adopt new ways of evading detection, so an enforcement activity may have a lower return in later years.
READER COMMENTS
Laron
Sep 7 2021 at 1:33am
Mahalos for this post. Can you (or anyone) help clarify this? You stated:
A standard number for the DWL from many taxes is 30 percent of the revenue raised. But the CBO analysis doesn’t even touch this; it’s just about auditing and collection costs.
So if the standard 30 percent applies, this amounts to an overall DWL of 70 percent of the amount of revenue raised.
I just don’t see the steps that increase the DWL from 30% to 70% of revenue raised. Any help?
Eric Rall
Sep 7 2021 at 10:23am
The other 40% is enforcement costs. $80B enforcement costs are 40% of the projected $200B in revenue.
David Henderson
Sep 7 2021 at 10:23am
$80 billion is 40% of $200 billion.
40% + 30% = 70%.
Alan Goldhammer
Sep 7 2021 at 7:31am
If the US moved to a simplified tax code, eliminated the corporate tax (or made it really low with no deductions) and put in place a VAT the IRS would be mostly redundant. For most Americans the government already knows how much one earns from various revenue streams because of various types of reporting. A number of Euro countries have simple tax filing where you just double check the government’s numbers and if there is nor error, you just accept it and bingo, your taxes are finished.
robc
Sep 7 2021 at 9:13am
I don’t see how the VAT connects to the rest of your point?
The simplification of the tax code and the elimination of the corporate tax both make sense in the context, but why add in a brand new tax?
Now, if you were advocating for move from income tax to consumption tax, sure, I get it. But I still wouldn’t do it as a VAT, I prefer my sales taxes to be on the bottom line and not hidden. And, from my understanding, VATs have a lot more compliance issues and complications that a straight sales tax – so you are back to needing a big IRS.
Evan Sherman
Sep 7 2021 at 10:29am
Regarding the last point:
“For most Americans the government already knows how much one earns from various revenue streams because of various types of reporting. A number of Euro countries have simple tax filing where you just double check the government’s numbers and if there is nor error, you just accept it and bingo, your taxes are finished.”
A lot depends on execution, which is to say a lot depends on competance (and, by extension, the public’s confidence in the IRS’s competance). To put it reductively: Americans tend to be more skeptical than Europeans of these kinds of policies not simply for abstract or ideological reasons (which also merit discussion, btw) but rather because of lower faith that our government will effectively and efficiently deliver on any proposed plan. I am probably generalizing and being reductive, but my understanding is that Europeans are more willing to trust their governments to do all kinds of things because they believe they can and will actually do them half-decently.
I say this as an American who has had to advocate for myself when my state and federal governemnts made a couple very obvious, avoidable errors on my tax return (i.e.simple errors that stemmed from incompetance). Definitely anecdotal, but still, it leaves a bad taste in my mouth and makes me less likely to trust that they would handle a pre-prepared tax return system competantly.
robc
Sep 7 2021 at 10:56am
I agree with you, but with significant simplification of the tax code, they probably could handle it properly.
Right now, for the vast majority of people, it is enter W2 info and take off appropriate standard deduction. That could be handled right now. Mine is more complicated than that. Yours is probably more complicated than that. But there could be lots of code simplification that would make it trivial to do.
The most recent estimate I can find (2017) is the tax code is about 2600 pages long (just the code, not regulations or rulings added on). Reduce that by 90% of so and we probably have a good starting point.
Evan Sherman
Sep 7 2021 at 11:58am
That is true: Simplification of the tax code itself would limit the risk of poor IRS or state level tax collection management/execution. (And it would be good on the merits, IMO, for other reasons.) As it happens, my anecdotal reference to issues with returns had to do with mis-entering / mis-reading my deductions on their end – deductions that, all things being equal, I would gladly trade for a lower base rate.
Comments are closed.