I just finished writing my review of Steven E. Rhoads’s excellent book titled The Economist’s View of the World. It’s excellent. In a longer than usual review, I didn’t have space to highlight his discussion of U.S. government subsidies to fossil fuels. We often hear how high they are. But Rhoads footnotes a Brookings study by Joseph Aldy that estimates those subsidies to total $41 billion over 10 years, for an average of just $4 billion a year. That’s not as high as we often see claimed.
All of the subsidies are implicit subsidies in the tax code. That is, they are provisions in the tax code that treat fossil fuels preferentially. The study was published in February 2013. So I’m fairly confident that updating for inflation since then would give a higher number.
On the the other hand, there’s an effect that goes the other way. The biggest single item (see his Table 5-1) is $13.9 billion over 10 years for oil drillers being able to expense, rather than depreciate, intangible drilling costs. But the 2017 tax cut permitted expensing for investments in short-lived assets such as machinery and equipment. So the preference for the oil industry suddenly fell. That would make the $13.9 billion for, say 2021, fall, possibly all the way to zero.
It is true that the expensing provision of the 2017 tax law was temporary. It starts to phase out this year and will be completely gone by 2026.
It might be useful for someone to do an update of Aldy’s study.
Note: There is an issue, especially for libertarians, about whether preferential tax treatment constitutes a subsidy. I’m always a little torn about this.
READER COMMENTS
Philo
Jan 16 2022 at 10:34pm
If and only if an entity is taxed *less than it should be*, it is receiving a tax subsidy.
Expensing for the cost of short-lived assets expires when? (Not 2016.)
David D Boaz
Jan 17 2022 at 9:22am
Define “should.”
Thomas Lee Hutcheson
Jan 17 2022 at 5:18am
There should be NO business tax at all, but I’d take the base line as profit tax that actually taxed returns to investment equally across sectors and activities. Expensing all investment would not favor one activity over another, but only “short term” investment (and what is short term about a hole in the ground?) does. Preferential taxation of unindexed capital gains can go either way. And don’t extractive industries still get to deduct percentage “depletion” in lieu of depreciation?
Matthias
Jan 18 2022 at 12:02am
Profit is rather nebulous and subject to accounting shenanigans.
To clarify: would you want to tax debt and equity the same?
At the moment, most jurisdictions allow to expense interest payments, but require dividends be paid with post tax money.
That skews corporate financing in favour of more leverage, ie more debt.
Thomas Lee Hutcheson
Jan 18 2022 at 6:26am
Not a problem if there is no taxation of business income. 🙂
But even so, what’s the problem with firms collectively having too much debt? Being “over” leveraged, they then have to behave “too” risk aversely to compensate? [I’ll be happy for anyone else to answer here. It’s a problem I’ve never understood.]
Dylan
Jan 17 2022 at 6:11am
David, a few weeks ago you linked to another study that looked at the global numbers for fossil fuel subsidies. You deleted the post because I think you had misinterpreted some numbers from the study, but if I remember correctly those put the global figures much, much higher. Most of those come from the developing world, and U.S. and EU numbers were quite a bit smaller, but I’m having trouble finding that study again. Any chance you could reshare the link if you remember the study I’m referring to?
Andrew_FL
Jan 17 2022 at 8:43am
Not the study from that post, but this one shows what you’re talking about
Alan Goldhammer
Jan 17 2022 at 9:21am
There is no doubt in my mind. Tax preferences are subsidies and they only increase with each new Congressional attempt to ‘fix’ the tax code. All of them should be eliminated and the tax code cleaned of all tax preferences. We recently moved into a condo and decided to take out a mortgage for part of the payment just to take advantage of the mortgage interest deduction. It made sense with interest rates as low as they were when we put in a contract. It will help balance off the limitations we face on SALT deduction. Are we taking advantage of the tax code, you bet. We do the same thing with charitable contributions making the straight from our IRA accounts so that tax burden is minimized.
If corporate taxes were either eliminated or moved to a very low level and replaced by a VAT, the kind of tax preferences that David writes about would be irrelevant.
robc
Jan 17 2022 at 12:21pm
I am a bit surprised that works. With a 10k SALT limitation and a $25k standard deduction, and very low interest rates, it would have to be a pretty expensive condo if only “part” allows you to itemize, unless you have lots of other itemizing items.
I normally itemize, but with selling our house in June, and our knew house not being done until this March, I am going to come up short this year.
And next year, with about 10 months of interest, its going to be close. I dont think interest and SALT will get me there, but other itemizations should push me just over the top.
BTW, I would like to see another large rise in the SD and eliminate itemizing altogether.
Thaomas
Jan 18 2022 at 6:40am
I’m quite in favor of favoring some kinds of consumption (charitable contributions, medical insurance and costs, home ownership vs renting is more debatable) over others. But whatever is going to be favored, let’s do it with a partial tax credit so the tax favoritism is the same for all income levels. A “deduction” for income not consumed (savings) on the other hand is consistent with a system that taxes consumption, not income. A SALT “deduction” also looks right in that those are resources that the taxpayer has not chosen to consume.
Thomas Lee Hutcheson
Jan 17 2022 at 9:30pm
Eliminating the intersectoral distortion in returns to investment is the main reason to get rid of business income taxes. True the are also a pretty crude form of taxing consumption compared to a progressive consumption tax or even the personal income tax, but I think that is a lesser order of magnitude distortion.
KevinDC
Jan 17 2022 at 11:27am
I tend to think so, although I usually try to be clear when using my language to describe it. Rather than saying “The government is subsidizing X through preferential tax treatment,” I usually phrase it as “The government is subsidizing X relative to Y through preferential tax treatment.”
To illustrate, imagine there are only two fruits in production, apples and oranges. In one scenario, the government subsidizes the orange industry by literally giving just giving them $1 billion dollars. In this case, I’d just say “the government is subsidizing orange production.” In the second scenario, the government doesn’t actually cut the orange industry a check. Instead, it given tax discounts and other benefits to the orange industry, which are in total worth $1 billion, but does not apply these policies to the apple industry. In that scenario, I think it’s fair to say something like “The government uses the tax code to subsidize the production of oranges relative to apples.” That’s not quite the same as saying “The government is giving subsidies to the orange industry” but I think the situation is described fairly.
And to echo Alan Goldhammer above regarding mortgages, I think it’s a fair description to say that the mortgage interest deduction amounts to using the tax code to subsidize homeowners relative to renters. I think this is a terrible policy, but so does pretty much every other economist in the world.
AMT
Jan 17 2022 at 12:53pm
So yes, it is a subsidy. This just adds extra detail to explain how the subsidy is transacted. Further, it shouldn’t be surprising this is why the government calls them “tax expenditures.”
KevinDC
Jan 17 2022 at 4:58pm
There’s a slight difference I had in mind that I think is being left out in your response. I was there’s a difference between the two statements I made, which is highlighted by the word you chose to put in bold. There’s a difference between “giving subsidies” to the orange industry, and using differential treatment in the tax code to “subsidize oranges relative to apples.” It’s the difference between a noun and a verb. I think the reason people can disagree so strenuously on whether things like preferential tax treatment counts as a subsidy is because it’s easy to conceptually conflate “subsidies” the noun, with “subsidize” the verb. Subsidizing is verb – it’s an effect, a thing done, not a thing given. Subsidizing an industry isn’t identical with giving an industry a subsidy. Giving an industry a subsidy results in subsidizing that industry, but not all subsidizing of an industry occurs by giving a subsidy. Or so it seems to me.
(Great, now I’ve said the words “subsidy” and “subsidize” in my head so many times that they sound weird. Semantic satiation!)
AMT
Jan 17 2022 at 7:35pm
I’d say the italicized portion isn’t correct, and that by definition anything “subsidized” received a subsidy, and that it’s irrelevant what form the subsidy took.
We know taxes and subsidies are opposites, so preferential tax treatment is, at the very least, the functional equivalent of a subsidy. So I think the only reason people can strenuously disagree is if they are economically illiterate, or they intentionally argue a meaningless semantic distinction is important, which is just plain pathetic. What decent economist would argue that this difference matters for determining whether an industry is effectively subsidized? None worth paying any attention to.
KevinDC
Jan 18 2022 at 10:52am
Yes, this is exactly what I was saying in my initial post. You can have the effect of a subsidizing, without actually granting a literal subsidy. Even if preferential tax treatment isn’t literally a subsidy, according to the standard definition of that term, (“A payment by the government to consumers or producers which makes the factor cost received by producers greater than the market price charged to consumers”), it still has the same effect as one. That’s was what I was saying the whole time. So we’ve both arrived at the same point there. Progress!
I disagree, however, about whether wondering over the distinction is economic illiteracy, or just shows someone is pathetic and not worth listening to, as you suggest. One can reasonably disagree with the idea “if they have the same outcome then they’re actually the same policy.” I think a person could reasonably hold that it only shows how “different policies can lead to the same outcome.” I don’t see anything contemptible in that. And if it’s coherent to say “different policies can lead to the same outcome”, it’s not the case that anything which leads to the same outcome must by definition be the same policy.
One of the Left’s favorite rhetorical flourishes depends on denying that distinction. In the same way that some would say “taking less money from the orange industry is literally the same thing as giving them a subsidy,” leftists love to insist that taxing people less is literally the same thing as giving them money. It usually takes the form of the cliched soundbite of “this tax reform bill is giving billions of dollars to the already well-off in America” etc. I don’t buy that description. I think there is an important difference in “actively giving” vs “taking less.” A leftist might respond by saying “Well, mathematically, a policy of ‘tax billionaires X amount, and then also give them checks for Z dollars’ works out the same as ‘lower their tax bill from X by Z dollars,’ therefore, this new tax law means the government is giving billionaires Z dollars,” I’d find that to be a pretty unpersuasive analysis on their part. The distinction still seems meaningful to me.
AMT
Jan 18 2022 at 2:13pm
Now please don’t make the mistake of trying to broaden the discussion beyond taxes and subsidies, which is what your quote does by just talking about “policy.” What you don’t seem to understand is that the only time this pathetic argument regarding subsidies comes up, is when discussing whether or not an industry or firm is subsidized. You are incorrect to think people are reasonably arguing “different policies can lead to the same outcome.” In reality, it’s an idiotic attempt to deny the outcome. They never say, “actually, they receive preferential tax treatment, which has an identical effect, but is technically something slightly different.” What they actually say is, “NO, they are NOT subsidized!!!!” They do everything possible to avoid admitting the identical effect. It’s meretricious garbage and everyone intelligent knows it because you can say that subsidies are negative taxes, or taxes are negative subsidies. There is no meaningful distinction. This is why I argued that preferential tax treatments ARE subsidies, and that “at the very least” everyone has to admit how they operate and understand why the counterargument people use, outright denial of the outcome, is worthless.
LOL. I mean, if you receive a larger tax refund due to preferential tax treatment, it is literally identical to receiving a subsidy.
And again, please stay on topic and don’t try to change the discussion to tax reforms that change overall tax rates, rather than preferential tax treatments and subsidies, which your last paragraph seems to be trying to do.
Henri Hein
Jan 17 2022 at 3:29pm
I don’t disagree with your language suggestion, but I don’t agree with this particular example. The mortgage interest deduction probably doesn’t benefit home owners at all. I expect home owners pay the exact value of the interest deduction in increased home prices. The deduction benefits developers, real estate agents, etc – anyone benefiting from higher house prices.
robc
Jan 17 2022 at 4:27pm
You are partially right, but not completely. Like a business cannot pass on all of a tax to the consumer (depending on the shape of the supply/demand curves), the same would apply to the subsidy. Much would go on to the developer, but the home owner would benefit some. If for no other reason than not everyone gets the mortgage interest deduction, so housing prices cant go up by the full value.
Max More
Jan 17 2022 at 11:48am
$4 billion per year is surprisingly low. It’s extremely low compared to subsidies for “renewables”. If the source I’m reading is right, Germany alone spent $220 billion over 10 years on renewable subsidies. (The result: 5% reduction in CO2.)
I agree with KevinDC in preferring the language: “The government is subsidizing X relative to Y through preferential tax treatment.” In my mind, the tax rate should be zero, so taxing less is not subsidizing in absolute terms. Rather, others are being preferentially taxed even more excessively.
Henri Hein
Jan 17 2022 at 3:25pm
That’s still $4 billion too much.
Land grants is also a subsidy. BLM at least allows private companies to extract oil and gas. The oil reserve also seems like an indirect subsidy. Keeping 600 million barrels off the market raise the price of oil.
alvincente
Jan 17 2022 at 8:57pm
BLM acts like a private landowner when it issues oil and gas leases. Like a private landowner, it receives a royalty for hydrocarbons produced. There’s no “land grant” and no subsidy.
Mark Barbieri
Jan 17 2022 at 7:47pm
Don’t forget that from 1975 until 2016, US oil companies were barred from selling oil to anyone outside the US.
Geoff Ryan
Jan 19 2022 at 9:55am
The bait and switch approach used in the cited study grossly overstates the value of “subsidies”. Thinking about immediate expensing, the right question is “how much value does the government give up by allowing immediate expensing instead of insisting on depreciation, which delays expensing”. The budget impact is a poor indicator of that.
Here’s why. In the first year, the oil company loses its deduction, but in the following years of course it gets back that same deduction, a little at a time. Suppose that the benefit to the government in the initial year is 100. There’s a cost to the government in the following years of 100. So properly accounted for, there’s no net “budgeting” benefit. But summing over 10 years, it looks fantastic! Because in year 10, the budget impact is 100 of benefit and all of the future costs to the government are neglected.
Just how exaggerated is the impact? Suppose oil companies give up 100 of immediate tax benefit each year and for simplicity would instead realize each year’s benefit over the next 10 years, straight line. After 10 years, the government’s budget impact is a whopping $550! Sounds impressive!
But what we should care about is value. How much value does the government achieve by delaying the tax benefits? A simple back of the envelope calculation using the current government bond yield curve is that the government has a net present value savings of delaying the first year’s $100 tax benefit of about $9. This is because rates are so low. How about over 10 years? The government benefit is obviously about $90 (due to the shape of the yield curve, a better estimate is $100).
That’s a lot of numbers, but the bottom line is this: the “budget” savings, $550 in our example, neglects a bunch of future costs to the government and, frankly, has nothing to do with true value. The value impact, $90 – 100 in our example, is much smaller.
I can’t do the math to get an exact number, but it’s clear that the budget impact is a poor way to talk about subsidies and almost certainly vastly overstate the true value of the “subsidy”.
Geoff Ryan
Jan 19 2022 at 10:14am
Let’s also talk for a moment about the second biggest “subsidy” in the cited study, the manufacturing tax deduction for oil and gas. I don’t know for certain, but I suspect that this is the section 199 credit, which is generally 9%. The credit is quite broad, covering even such activities as a making hamburgers and roasting coffee. As former US tax official John Harrington explained, “if you are in the Down Jones industrial (average), and you are not taking this deduction, there must be something wrong”.
So, while it’s technically a subsidy, I guess, it is not a subsidy that works to the advantage of the oil industry relative to other industries. An even handed approach would be to either revoke the subsidy for everyone, or to leave in place for everyone. Showing it in the table as a subsidy to the oil and gas industry is misleading at best.
But wait! There’s more! The 9% manufacturing credit applied generally, but the oil and gas industry was singled out, and received a LESSER deduction of only 6%. In other words, on a relative basis, the oil and gas industry was punished. So an even handed approach would be to show that the oil and gas industry didn’t receive a benefit of $$11.6 Billion relative to the rest of the economy, it was punished to the tune of $5.8 billion.
So looking at just the biggest two items in the study’s table, it seems clear that the first number is vastly overstated, and on a fair and even handed basis the second biggest subsidy isn’t a subsidy at all!
Wow! I wonder how inflated all the other values are?
Geoff Ryan
Jan 19 2022 at 11:47am
One final small comment. My back of the envelope calculation used straight line depreciation to calculate the economic benefit to the government, but IIRC the US system is an accelerated deprecation system. Thus my estimate overstates the benefit to the government of eliminating immediate expensing. Further, fracking wells have a very short economic life with production heavily weighted to the first few years. As unconventional (fracking) rises as a proportion of US oil and gas investment, the economic cost to the government of immediate expensing relative to expensing over the economic life of oil and gas investments falls.
Comments are closed.