This is a topic about which I'm very passionate. I've written my congressional representatives several times. I've posted on LinkedIn. I'm raising awareness with my close colleagues. I'm trying to beat the drum as much as I can.
There is a minor win that was just advanced in the House yesterday[0] to delay US onshore amortization through 2025 via The Tax Relief for American Families and Workers Act of 2024[1]. It doesn't touch offshore amortization.
It's not enough. I was geared up to start a business this year; this, specifically, has put the brakes on it. I've written as much to my congressional representatives. I'm doing research into what it would take to incorporate in another country, and the implications therein, if it comes to that.
> It's not enough. I was geared up to start a business this year; this, specifically, has put the brakes on it.
I've read a little about this, but not a lot. How has this put sufficient drag on starting a business that it's impractical to go from zero to one? Is there an aspect of the business model that's particularly impacted?
Your salary costs for all programming and other development can only be written off over a five year period instead of the year you incur them. That means if you are breaking even normally with 1MM of developer costs, the government will tax you as though you made 800k. Because it says 80% of the code they write is a long term investment and therefore needs to be written off over time like a company car or a drill press.
For mature companies, it all evens out, because you write off the costs over 5 years and it's just a blip. You can just borrow that money as a corporate bond, creating a real but manageable cost. For growing companies, it's forcing you to set aside money now (in the form of a tax deduction) you won't get back until year 5. So your burn rate is going up. And, unlike a mature company where the cost decreases over time (by year 5 it was a one time blip), that's only to the degree your salary costs have been stable for five years. Every time you scale up engineering staff, it puts you more in the hole for another five years.
It’s not five years. It’s whatever the useful lifespan of the software. Web dev we usually did 2 years since that was the average lifespan of a web page.
For public companies, they are incentivized to amortize over longer period because it hides expenses and can boost earnings that boosts stock price.
Talking about software development which can be classified r&d for tax credits but it can also be capitalized like any other assets.
This pdf has handy chart to determine how to determine software development should be treat for accounting purposes. Grant Thornton is the external auditor for many large companies.
No. TCJA requires all software development costs to be treated as r&d, and capitalized over 5 years (or 15 years). Guidance from pre-2022 no longer applies.
My reading of this document is that certain software development activity (e.g., “corrective maintenance to debug, diagnose, and fix programming errors”) is not SRE.
This seems to contradict your statement that “all software development costs [are] to be treated as r&d,” but my experience is that you know what you’re talking about. What am I missing here?
I would suggest reading notice 2023-63. That provision is very limited, and may not be available to you at all depending on the software you're developing, and/or the stage of the development that those steps occur in.
For example, for a new feature: if you plan (SRE), design the interface (SRE), write the feature (SRE), run it through QA (SRE), and then discover a bug and correct it -- that's still SRE.
If you put the software into prod, then discover a bug, then fix it (without improving performance or adding any functionality), then it might not be SRE.
But if you sell software, and you sell or install a release to a customer, and they (or you, under a support contract) discover a bug and fix it... but then you include the fix in your next release, probably SRE.
Or, if you put the software into prod, discover it breaks with a large data set, and you fix it by improving the performance of that section of code, probably SRE.
The expenses falling under that provision aren't going to make a significant change to the impact of TCJA on software development.
Third to last paragraph: if I sell software, a third party bug bounty type guy finds an exploit, notifies me, I patch the vuln, and release the patch as a hot fix: you figure all of that is SRE, probably?
(Disclaimer: not seeking tax advice or legal advice, we’re just two dudes casually discussing section 174 like normal people do all the time).
That notice was a helpful read. I think what I may have been missing was section 5 of Rev. Proc. 2000-50, whereunder non-SRE software dev was also afforded some similar protection. I’ll read that after work but I’d still love you to answer the foregoing.
I re-read the section of the notice that scenario would apply to, and I actually think its pretty clear that is not SRE. Correcting defects discovered after the software is put into production, or discovered in released version of software, and not considered SRE. See section 5.03(5)(b) of the notice. Pre-release bug fixes are still SRE though.
Definitely ask your CPA though. I'm not an accountant.
> 26 US Code Section 174, paragraph D - Treatment upon disposition, retirement, or abandonment
If any property with respect to which specified research or experimental
expenditures are paid or incurred is disposed, retired, or abandoned
during the period during which such expenditures are allowed as an amortization
deduction under this section, no deduction shall be allowed with respect to
such expenditures on account of such disposition, retirement, or abandonment
and such amortization deduction shall continue with respect to such expenditures.
This means something different than adrr is asking about. The IRS has depreciation schedules for different asset classes. For example, trucks are 5 years; real estate is 39 years; but you can under some circumstances use an Alternative Depreciation Schedule (ADS).. the depreciation schedule should match the usable lifetime of the asset.
But if the typical lifetime is 5 years, but you use an ADS of 2 years... you are not disposing or abandoning the asset if you keep it for 2 years. 2 years is the expected life time of the asset, and at the end of year 2, the asset has a value of $0.
If you depreciate over 5 years, but then on year 2 decide you don't need the asset anymore, then you'll dispose of it. The asset is valued at 3/5ths of the original price. The paragraph you're quoting applies to this scenario.
ADS doesn't apply here though, because TCJA requires 5 years for domestic and 15 years for foreign research.
I am aware of that part, yeah. It's a shitty change, one hundred percent, and I'm glad to see folks like Wyden trying to unwind it. At the same time, I've owned small software businesses before, and I've started/not started new ones based on whether I thought it would work out. Where I am unclear is what sorts of business plans, on day zero, would be suddenly nonviable when it was viable before, so suddenly overwhelmingly difficult, as to cause such discouragement.
Is there a good chance of recategorization pushing a very marginal business off the cliff? Sure--that Twitter thread, where there are enough details to read in, has a lot of examples of marginal businesses having trouble (mostly because of not being able to plan in the change, which really sucks). But you don't have that on day zero, and going in with the aim to be that very marginal business is probably not the best of ideas. It's not a day-zero problem, and I am struggling to see where a small business with a business plan that was previously worth executing on is now not worth it because of this change.
It fundamentally changes the cost structure and investment risk profile of business plans without changing anything about the intrinsic economics of the business by requiring much more capital to achieve the same outcome. A perfectly reasonable business plan can suddenly become non-viable if there is a huge new overhead to doing business. Suddenly needing to pay $1M to the IRS on "profit" for a company that is barely making money is rather large change to the financial assumptions that make the business viable.
Affected small businesses suddenly need to increase revenue or cut costs by 20% just to keep their business solvent. Most small businesses do not have the structural elasticity or capital reserves to absorb that, nor do many business plans. In the very long term it notionally all evens out but most small businesses don't survive that long and these large new costs of doing business will reduce the survival rate even further.
> imagine in year 1 you grossed 100k, spent 200k on salaries, but the new irs rules say you owe taxes on 60k of profits.
For a clearer picture on this it’s helpful to include the actual tax amount. With a corporate rate of 21%, the tax would be about $12K.
Now that’s not zero, but it better shows the actual cost born in year one under this plan.
The later years are important as well. As profits continue to flow in, the rest of the cost can be deducted from it. So it doesn’t disappear.
Though you do lose a bit from the nature of present nominal being inherently worth more than future nominal amount. With todays higher rates, that’s an even larger factor.
Right, this is exactly why I'm confused about the all-is-lost framing of this. $12K isn't nothing. It's a pretty shitty thing to have drop on you and I think it should revert. But if you're starting a software business that isn't a solo shop--so you have hundreds of thousands of dollars in payroll--I would expect a $12K difference to not be the needle-mover between "start a business" and "don't"; we're not talking a hot dog stand here.
I understand where you’re coming from but the thing to understand is: most startups are incredibly marginal to begin with. If you tilt the field such that the entire distribution is now ~5-6% lower expected value for any given outcome, you can easily wipe out 80%+ of the startups that might’ve been “worth trying” but don’t make any sense now on a risk-adjusted basis. The entire asset class can become unfundable because the same money taking the same level of risk simply generates more returns elsewhere
Are we talking "startups" or "small businesses" here? I could see this blowing up startups for sure, but the conversation was about small businesses and I'm not sure I see the connection there. In my conception of the universe, you're not rolling out multiple hundreds of thousands of salary when you're hanging out your small-business shingle. You might grow to it, but that'll be over a long period of time.
Having to amortize salaries mean that a freshly starting company has to include any salaries paid as (software devs salaries)4/5(corporate tax rate) extra thing to budget - if they don't have access to necessary reserves up front, it might be enough to turn a marginal business out from starting at all.
Tech startups are usually “let’s take a gamble” not “let’s pencil in 30% tax deductions.” I’m all for not paying taxes but it’s not typically part of the calculus before starting something.
Which is why so many are having issues. Having to come up with a 30-60k per engineer on top of your expected salary burn rate can have serious impacts to a startup runway.
Michele Hansen is doing an incredible job here and should be recognized for it. Latest news was sent yesterday:
Republicans and Democrats finally struck a tax deal that includes a partial fix for Section 174. It includes expanding the Child Tax Credit, a key Democratic priority, with a handful of business tax issues where were Republican priorities...
Not sure if there is a public copy of that email, but the core info is on the site, it has a list of ways to get your representative's attention and a script to follow. Please call.
You should note that these sorts of tax law changes are largely a revenue game: Someone wants to cut some other tax, so they find a place to raise taxes elsewhere without thinking about the consequences. But also, the vast majority of people employed by R&D-heavy companies are Democrats, so it's hard to impute motive one way or the other.
Independent/unaffiliated voters are now 47% of the electorate in the US so that last statement probably doesn't hold water. Are educated voters, regardless of party affiliation, more likely to vote for serious candidates that focus on issue resolution over culture signaling? Sometimes, but that no longer has much to do with party affiliation.
I don't think that's the definition usually used for "independent" when talking about voter affiliation. US politics has a long history of people telling pollsters (and filling out voter affiliation info in actual elections) that they aren't affiliated with a specific party because political parties have generally historically been seen pretty negatively in the US. In reality, many of these self-identified "independent" voters have voting records that are heavily skewed (if not straight-ticket) to one party or another. I don't have actual numbers on what % of independents this is (I too am interested in the parent's question on this matter), but it's definitely not a "definitionally, none" answer.
I'd argue that voting predominantly for one party or another doesn't identify the unaffiliated individual as a member of the party who's candidates they vote for. It absolutely doesn't identify the whole unaffiliated voting block as having any secret party affiliation. Telling the whole party system to take a hike is an active choice and it happens in counties that only allow a single political party as often as it happens in countries like the US where there are only really two choices.
A lot of politics consists of quietly creating problems and then loudly coming in with tons of press coverage to solve it later so you look good and get reelected.
I wish there was a good way to prevent such behavior.
It's was part of the very same tax package and implemented for the same reason: It made the trump tax cuts look economically more reasonable by sticking in time delayed increases on tax burdens that would land on disproportionately on democrats. Yet democrats were unable to undo it because dem platform programs need the funding.
On what basis will it be different here? Dems are almost completely disengaged on the bill intended to fix this, almost all the activity for and against appears to be republicans.
Addicted's confident statement "Even when this fix is passed, it will be voted for unanimously by Democrats" seems misplaced and unlikely. If that were likely we would have seen broad dems support on fixing the SALT tax limits and we just didn't.
The strong bipartisan showing in the House Ways and Means Committee adds more momentum to the proposed changes, which include allowing the immediate expensing of research and development costs.
The challenge at the moment is to get Speaker Johnson to put it on the floor and give him confidence that he has the votes. Critical that everyone contacts their Representative next week to ask them to ask leadership to do so, especially if their Representative is a Republican.
Going on a tangent here, but I always hated this idea of reading off a script to try and lobby politicians. If I were the politician, I would ignore all such communiques as indoctrinated spam.
Totally fair! I provided a script because many people have never contacted Congress before and might not know what to say, so it’s a way to make them more comfortable with the idea. They don’t necessarily have to use it.
Now for tweets, it’s important that each tweet is unique. Otherwise it’s clearly coordinated/automated. But as a casual scroll of Twitter shows, many people hate comfortable expressing themselves through that medium.
Also, I kinda wrote a whole book with scripts (for customer interviewing)… if I didn’t provide scripts, it just wouldn’t be true to my work!
Typically staff answer the phone. They’re making tally marks. If the script tells the staffer what the issue is and the caller’s position so an appropriate tally can be counted, it does the job.
This is going to wipe out a lot of small businesses, including innovative software dev startups.
Here’s a simplified example of this works: Let’s say you’re a four person software dev startup. Everybody is making, say, $125K to get by. That’s $500K in salary expense which normally you can write off as expenses against revenue/funding. For this example, let’s say somehow you also generated $500K in revenue/funding, i.e. you just broke even.
Currently, you would (of course) owe zero taxes. Under the new tax rules, you couldn’t write off those R&D salaries as expenses, only amortize them over 5 years. That is, your salary expenses for this year is only ~ $100K, and this you made a $400K ‘profit’ (!!) on which you owe taxes ($100K).
Further, R&D is at best a bet, as any of us who has ran a technology business well know. It’s not at all clear one will be able to take the research to market and monetize the R&D investment. Doing that is called founding a successful company.
And if the investment pays off, so that the R&D creates revenue in future years after it's already been paid for, then it will result in actual taxable profit in those future years.
It seems like the IRS is trying to speculatively tax unrealized future profits here, and that's pretty unconscionable.
This is exactly what they are doing. If this scheme sticks, it will reduce R&D investment. It certainly disincentivizes me from investing in such efforts.
We once started a revolution over unfair taxation. I sadly think the same will be necessary to fix the status quo, as system has rotted past the point of meaningful reform.
Allowing the federal government to implement an income tax was one of the most destructive changes to the constitution we've ever made, and it's only gotten worse and worse over the decades. What was originally just supposed to be another tool to generate revenue has become an end in itself at best, an a lever to manipulate society at worst.
On top of that, the necessity of shoehorning everyone's economic activity into prescriptive taxonomies in order to implement this taxation has drastically diminished innovation on the margin, and has created unintended consequences that significantly distort market incentives and diminish our ability to adapt to change, with the situation we're discussing here being a prime example.
I'm worried that you're right; that it's beyond repair, and that we may have to live through a systemic collapse before things can get better.
It is salary, but it goes not into operational expense (that applies directly creating the things you sell right away) but into building a capital asset that helps you generate future revenue and can depreciate over a number of years.
It's just as if you pay the salaries of some construction workers to build you a new factory - those salaries are part of the capital investment for creating that factory (and depreciated as that asset), not your operating expenses.
As a SMB, if your salary expense is $500k and your revenue is $500k, you are absolutely struggling, no matter the tax situation. In this case you "just" need another $100k in revenue. After all, you are throwing out theoretical revenue numbers anyway. So it just moves the needle somewhat for you to still break even. You could game it just a little if there's revenue at end of year that you can book in the following quarter instead -- companies do the reverse of this all the time. Perhaps you can shift some of the expenses as well.
Note that in year 2, you get 20% of the year 2 salary expense, plus 20% from year 1. So the impact is less. By year 5 you are "caught up".
Companies don’t do “this all the time”. In your hypothetical example, you”d have to essentially forge the quarterly 941 reports to shift the salary R&D expenses. Also ‘gaming’ which quarter revenue occurred is technically tax fraud, certainly if one is on an accrual basis. And it doesn’t mean one is ‘caught up’ by year five. Any time a company increases R&D expenditures, it will have an impact.
Have you never bought anything? There is always a push at end of quarter and end of fiscal year, to sign the agreement NOW. so that it can be booked in that quarter. even though delivery won't happen until next ...
incentivized by quota but the incentive is there to get it on the books.
there is no forgery involved.
Yes, of course as you increase spend you increase the impact (in the year that you actually spent), but the context here is specifically very early SMBs, and more so, ones that would only be running break even under previous rules.
This is wrong because you are speaking on an accrual basis but you have to consider cash. The cash (salaries and taxes) is paid out in year 1 regardless of what year you try to book it in your p&l statements
i'm suggesting that as an early startup, in year 1 where this has the most impact (you aren't now bringing in previous years expense), your employees may be flexible enough that you pay out the cash 31 days delayed (example) instead of paying anything in December.
obviously if you pay salary in December, you paid in December and you have to book it that way.
I think I am being fair. The numbers $500k/$500k were picked arbitrarily to demonstrate how it would kill a business. If you were at $500k/$400k you would be similarly dead. "back then", $500k/$500k was enough to stay afloat. Now you "just" need to be at $500k/$600k, that's all. ie, your revenue needs to be somewhat higher (20% I guess? too lazy to math it) to pay the taxes. Assuming the business stays afloat, you'll eventually break even on the taxes, it's not forever lost like paying AMT for stock options that never become liquid, or RSUs that go well underwater before a lockup expires.
It does suck that a quirk in accounting means you were profitable on the books, no question there. But it "just" means you can't run so close to the bone. I don't think it's the world ender everyone is making it out to be. It means you need more operating capital up front.
More detail with math here. Lawmakers are discussing delaying these changes for a few years and/or allowing deductions for domestic employees, but all depends on signing a budget, which is never certain.
So far as I can find the last time Congress bothered to pass a budget was 2016 with the prior one in 2010.
The government has been funding itself instead with “continuing resolutions” which pretty much just continue spending as the prior year modulo marginal changes. Incidentally this is why federal deficits have exploded since 2010: the financial crisis “one time” trillion dollar stimulus has been continued every year since.
It was about three months late, but they passed it. This year's is currently at least about four months late, and the federal government is currently operating under a continuing resolution that will expire in March. However, full-year continuing resolutions are rare: the budget bills are usually passed, but passed late.
Not how deductions work. Say 20% corporate tax, a 100% deduction (which is the normal business expense) means you don't pay any tax. A 135% deduction means you'd get to expense 35% more, avoiding those 20% for a "total saving" of 7% compared to a normal expense.
A credit does, but a deduction does not. I'm not sure where the previous commenter is getting this information from, since I can't find any documentation of a 135% tax deduction for R&D in Switzerland anywhere online, but if it is a deduction and not a credit, it means that it just offsets taxes on other income.
I hear this a lot and it's difficult to reconcile with my experience. Every US company I've worked at (and that's always been in at-will states) has not made it easy to get rid of people. Even people with woefully bad records of losing money every year and having multiple harassment complaints filed against them were kept for nearly a decade. With one exception (I personally got fired from a tiny startup because I refused to commit timesheet fraud for the CEO), the stories I've heard of the lengths that European companies have to go through to fire someone sound exactly the same to the processes I've seen at all of my employers.
You describe a scenario where management didn't want to fire someone - that's why it was harder. In the U.S. only two things get in the way: A) venial corruption B) worrying about unemployment insurance (that's why HR makes you do paperwork documenting an issue).
In many European countries you have to file a ton of paperwork and justify it: ex. at Google, they're still working through _January 2023_ layoffs because you have to work with the government itself and there isn't a good* financial reason for it
* by European standards. "we need stonk to go up" doesn't fly if you're massively profitable
Ive worked at a lot of different earlier stage software companies in the US and we've always fired very quickly, especially if there was harassment, but also just for low performance. Were you working at bigger companies? (aside from the tiny startup where the ceo wanted you to commit fraud) This hasn't been my experience at all.
> Every US company I've worked at (and that's always been in at-will states) has not made it easy to get rid of people.
That's an internal choice they do, to avoid having a reputation of a company that fires people any second (but then you have companies like netflix which take pride in having that reputation, but make up for it by paying more).
However, it's very different from European companies where these processes are (often) driven by laws. In the US there are no employee protection laws (aside from protected classes) so even if the company has a rigorous internal process, they could at any second override it if someone high up says so and you'll be fired in the blink of an eye.
It's very easy to fire in Switzerland. The notice periods are usually longer than in North America, but everyone having unemployment insurance where they are paid ~80% of their salary for up to 2 years makes is not such a big deal.
Can you say more about unemployment insurance? I haven't heard of this [a us worker] and honestly I also wouldn't mind being more aggressive with my career if I can guarantee ~80% of my income for 2 years should I lose a job.
You have a mandatory deduction on your salary (2.2%, a bit less effectively if you make over 150k). You need to have contributed for at least a year in the past 2 years before you're eligible.
You get 70% of your salary (or 80% if you have children under 25) for two years, capped at 70% (or 80%) of 150k.
There are a lot more exceptions, special cases and so on, but that's the gist of it.
It's state by state and different states have vastly different rules.
Here in Virginia, the max, regardless of how much you made, is $378/wk, for a max of 12 weeks.
You can't claim it if you're also receiving a severance, and you also have to record at least 4 job applications each week, but the documentation required needs to include information like the hiring manager's full contact information, which usually means the company needs to have replied to your application within that week.
I got laid off back at the end of April last year. I got a month of severance, so I couldn't claim UC in May. In June, I was able to do some online sleuthing to figure it out for a few applications out of the dozens I was making in a week, but there were some weeks I wasn't able to scrounge together even 4. I ended up with 2 UC checks for a total of $756 gross (yes, had to pay taxes on it). I don't remember exactly how much, but I do remember I calculated it was less than 20% my original take home pay for a month.
Luckily, by the end of June I had a good line on a job and started in July. I got lucky that we could bridge a month of basically "no" income from me. I can't imagine what it would be like for a single-income family living here in one of the most expensive areas of the country.
It's a kind of a tax - computed from your income. Similar to health insurance (in Europe). There are caveats like "you must be actively looking for a job" and "you weren't fired for an offense".
You are mixing up EU and Switzerland, employment laws are very different (and each EU state has its own, but generally much more protective of employees than Swiss ones).
One of the reasons Google has long term big center in Zurich, if grass would be greener (since cheaper it is) in say Germany or Austria they would build there
> One of the reasons Google has long term big center in Zurich, if grass would be greener (since cheaper it is) in say Germany or Austria they would build there
I don't think Google has an office in Zurich because it's cheap. It's mostly due to a lot of talent available (ETHZ, EFPL, etc).
You MUST spread it over 15 years, which is brutal for most companies and will mean no longer hiring any foreign R&D or software development contractors.
My feeling is that an EoR (employer of record) like remote.com might be enough for everyone to be able to avoid R&D capitalization. The research happens in Canada, by a Canadian employee of a Canadian company. You pay an American company for outsourcing human resources. The American company pays the Canadian company for human resources services.
Then again, if you're paying another company for outsourced human resouces, but you have an IP assignment clause as part of that, maybe you would need to claim it as your R&D expense
So this year if I spend $150k on foreign research (which includes ANY software development), and then I also earn $150k in revenue: despite me having $0 in the bank, I will only be able to deduct 1/15 of that, or $10k. In other words, I’ll be taxed as if I made $140k of profit, despite me not having any actual money left over.
You can see how if this was 5 years, then I could deduct 1/5 and I would be taxed on $120k profit, which is still bad, but not nearly as bad.
I don't think it will stop foreign hiring for R&D, since the cost differential is often greater than the tax obligation would be.
E.g. if you pay $150K for local research, you expense $15K (10%) the first year (and 30K the subsequent year). You pay taxes on $135K of 'profit'. Let's say that's $45K (I have no idea what's realistic here.
Alternately you pay $130K for a dev from Canada, expense $4,333 (1/30) the first year, and pay tax on the remaining 'profit' of $125,666. Even after admin costs you're coming out ahead
Just as an aside - as a general accounting principle, no, wages are not always deducted.
The easy example is a car company, like Ford. If they buy a car factory, that is a capital asset, and the cost needs to be amortized over x years. If they decide to instead BUILD a car factory...they still end up with a capital asset, and the costs (including wages) need to be amortized over x years.
In most cases this is what companies want - they'll have revenues over x years and matching costs over x years is generally better for everyone.
This isn’t like a loan where the longer the term the smaller the payments. It’s the reverse.
You essentially pay taxes now on income, and can’t deduct costs for 5 or 15 years. So it’s kind of like pre-paying taxes and not getting the money back for 5/15 years. Say that you need to go borrow cash to cover the shortfall. Is it cheaper to borrow money for 5 or 15 years?
As far as I understand it, it’s normal to deduct most all kinds of payroll as an operating expense, and historically that’s included software developers too. The way it was explained to me, the tax man gets his bite when the people receiving the paychecks pay their own income taxes.
The recent changes mean you can still do that for most staff EXCEPT developers, even if the devs are doing operational work instead of work that feels more conventionally like R&D. So you have to come up with a bundle of cash now to pay tax on most of the developers’ salaries, even though they’ll give it back to you over 5-15 years.
Essentially you making a free loan to the government for a decade or whatever, except the money’s probably not free to you.
Of course I can think of situations where the development effort really was more R&D than operational, and the revenue stream matched: the first few years operated at a loss already, and the deductions might have more been useful in 5 years when the revenues were flowing in from a mature product. But I think they might have ways to carry forward losses to future tax years or something to deal with situations like that?
So if you run a McDonalds. Your cash flows is money coming in for finished burgers. Your expenses are workers salaries and to buy meat. Let's say your very simple business summary is:
* Total Revenue: $1,000,000
* Cost of food: -$200,000
* Employee Salary: -$600,000
* (potential profit): $200,000
Assuming you can deduct the cost of food and employee salary to sling your burgers.. you make a 200k profit, and pay taxes on 200k.
But now let's say you can't deduct employee salary. You now pay taxes on $800,000 of income despite only having $200,000 of income. Depending on tax rates etc. you might end up with $0 in your pocket, despite having a successful business.
Now replace McDonalds with bootstrapped startup, food cost with AWS bill, and keep employee cost. This is the real situation many small SaaS or other software companies are currently in.
Maybe unpopular but I assume this is just what is going to happen. I remember awhile back when I was visiting Ireland one of the tours mentioned that modern whiskey and beer in Ireland came about because of taxes. I wish I could remember more details but the story went along the lines of each time a new tax on some type of alcohol happened the producers would reclassify their beverage or change the method which produced it to avoid the new taxes.
I assume this will be the same in the US, software R&D is now some other title with less taxes. When the small companies do it the IRS isn’t going to care but then big ones will do it and the tax law cat and mouse game will continue.
The difficulty with this is that §174 counts "all such costs incident to the development or improvement of a product." This makes it hard for you to reclassify the researcher role as any kind of software development role because software development is still R&D. The fact that it includes payments to third parties for R&D also means you can't even just split your software dev house into a nonprofit or something. What can you do? Claim prompting an LLM is not software development even if it produces code? It's just search then, right?
Yeah, that's what I'm guessing many smaller companies will do. Does the developer talk to customers or do customer success/support? Great, they are now "Sales Engineers" for tax purposes.
(Note: not recommending to do that, but guessing that is a natural byproduct of an unfair tax code).
To be fair, accelerated R&D amortization (immediate full expensing in the year the expense was incurred) is a tax loophole. Essentially, the default tax treatment of expenses is basically to take the expense same as you would treat it under GAAP, but some people (I am one of them BTW) think that we should put a finger on the scales for the case of legitimate R&D.
Now though I happen to think accelerating it is a good idea, everybody thinks their particular loophole is a good idea. To say that removing the special treatment is a policy mistake is a reasonable position to take, though opponents have a reasonable position as well (as I said I'm in favor of the special benefit). But to call this change unfair is, IMHO, unreasonable.
It's also bogus to plead ignorance as the twitter poster did: "as they'd never had to amortize software development before, and didn't think of the work they do as R&D." Their accountants sure did, because otherwise it would not have qualified for the R&D exemption. And their accountant would have to tell them what to do to make it qualify.
I don’t understand why businesses can’t just say that their engineering department is a cost center (COGS) instead of classifying engineer salaries as R&D expenditures.
I guess one downside would be not qualifying for the R&D tax credit.
The majority of software engineering is the equivalent of janitorial work… keeping servers online, fixing bugs, maintaining services, upgrading and refactoring code, etc.
It’s difficult for me to tell how much of this issue is just people making a fuss about the literal interpretation of the tax code compared to how it will, in practice, impact their business.
Nearly all accountants try to operate in the gray, in that accountants know how to tweak the numbers just enough to not get in big trouble with the IRS but enough that the business can slide through various loopholes. I have a feeling most accountants would just say “reclassify your engineers as maintenance workers, COGS” as a solution.
> I don’t understand why businesses can’t just say that their engineering department is a cost center (COGS) instead of classifying engineer salaries as R&D expenditures.
I spent a bunch of time on this last week with my experts-for-hire, as well as reading IRS guidance and 3rd party analyses. My understanding is:
- Previously you could decide whether to capitalize/amortize your R&D expenses or not. Now you must capitalize/amortize.
- Previously, you could choose to take the R&D credit for R&D activities, or not, regardless of whether you capitalized/amortized. That is still true.
- Previously, software development was only considered an R&D activity in certain circumstances. Now, the IRS has "clarified" that they consider the process of software development to be so similar to the process of traditional R&D that it should nearly always be considered an R&D activity and therefore should be capitalized/amortized.
One thing that's been frustrating is how, in discussions about this issue, software development is being spoken of nearly exclusively in the context of businesses developing software for themselves, either for internal use or resale. Left universally unmentioned are all the contractors, development shops, etc. who are developing software on a work-for-hire basis. How these companies should classify their engineers' work is not particularly clarified by the IRS guidance, but my understanding is that a consensus of "big" accounting firms is that these salaries should continue to be deducted as they were before.
I’ve many years of experience over a variety of companies and none of my jobs were janitorial work. You’re describing SRE or intern work. Many many many Software engineers build things, especially in startups which is the focus of the discussion.
> I don’t understand why businesses can’t just say that their engineering department is a cost center (COGS) instead of classifying engineer salaries as R&D expenditures.
They used to be able to. Then Trump and the Republicans came into power and changed it in 2017; the first year to be effective is 2022.
Before the 2017 change you could and should have if they were not engaged in actual R&D. But people used to lump them into R&D anyway because of the preferential tax treatment (you also used to get special treatment for NOL, which you can't any more).
Software development doesn't generally follow GAAP's model because unless you're IBM or Oracle the same people sometimes fix bugs and sometimes develop new features.
The GAAP rules aren't insane: if your company buys a HQ building it's assumed to last 40 years. And since each year you "get some value" (e.g. you don't pay rent to someone else) you split the value of that purchase over 40 years. If you buy a computer you amortize it over 3 years because you're likely to "use it up" and replace it after that.
And so if you're a car mfr you might have some people developing a new automatic transmission. You expect to then put those things into new cars for some time to come. So GAAP says the R&D that went into the new transmission is amortized while the cost of making one and sticking it into a car you sell is indeed COGS.
If you're an early stage pharma startup almost all you do is R&D for quite a long time. For a software startup you can say the same thing: even if you ship your MVP, every bug you fix is turning it into the "real thing" (unlike Oracle, above, who can split maintenance and development into two buckets).
So I favor accelerated recognition of R&D costs for startups. Not sure about established companies.
All of payroll is lumped into 1 line item when submitted to the IRS. (I’m only somewhat simplifying for sake of argument)
We would have to go out of our way to split out R&D.
When you win $500 at a casino, do you go out of your way to claim the gain on your taxes? Absolutely not. Are you supposed to? Absolutely, but you don’t. And the IRS doesn’t much care.
I feel like we’re talking past each other. The vast majority of companies won’t be paying more taxes because of this rule change.
Do you not keep your books according to GAAP either?
If you make up your own accounting rules and it seems to work for you, well, more power to you.
I like to have a clear understanding of where my business is going and the GAAP rules are pretty reasonable and align with other companies and what they do.
We do. We also have professional accountants who manage all this for us.
It’s interesting that HN is more concerned about the rules changes than the person filing our company’s taxes (or our CFO). This is a topic that is basically glossed over as a non-event between us and those professionals.
(Unlike the wayfair ruling which sent everyone, including our accountants, up in arms about sales tax and how to decide nexus)
Maybe I have a bad accountant and CFO? Or HN is extrapolating this to mean more than it does.
> It’s interesting that HN is more concerned about the rules changes than the person filing our company’s taxes (or our CFO). This is a topic that is basically glossed over as a non-event between us and those professionals.
You are the one who originally asked why engineers can't be considered a cost center. There is a reason; the R&D tax credit.
> Maybe I have a bad accountant and CFO? Or HN is extrapolating this to mean more than it does.
I am guessing you just do not qualify for the R&D tax credit or it just not worth the paperwork at your scale. For larger companies (and profitable, fast-growing startups that do a lot of the 'develop' part of R&D, which are rare these days, I guess) there is a benefit.
Considering even the tax foundation claims would bring in less than 20k jobs, maybe Congress made the right decision to bring in amortization even if I disagree with what it funds. I do not have numbers for how much it brings it to the federal government but the credit was already known for being difficult for small businesses to use.
i don't know enough about tax codes and how business works to comment on the actual policy, but it sure does seem to me that changes that result in immediate and massive new liabilities are unfair.
> changes that result in immediate and massive new liabilities are unfair.
Congress knows this: the law was passed in 2017 but only took effect for tax year 2022. This very issue was widely discussed at the time.
That was five years to figure out what to do and your CFO (or at the very least your tax accountant) should have been warning you. I mean, if your tax accountant doesn't know the tax law that's a bad thing.
I think the big problem with this is just the accounting burden: now even salaried programmers have to track hours between greenfield and maintenance development.
I suspect you could use your infrastructure ti give you a pretty good idea of where the split is as you have tools like your bug tracing system, different branches on repos etc to figure out adequately how much time goes where. Much better than time cards unless you’re a government contractor.
I respectfully disagree. By these arguments, literally every deduction from revenue to calculate taxable income is a "loophole". How far does that go?
What's more absurd than life itself is that you can deduct 100% of a 6000GVWR truck which you financed for 7 years, but my engineers salaries aren't deductible because I'm "building some product" so that's "development".
> Any accidental policy that disincentivizes R&D is stupid
The problem with section 174 is not that it disincentivizes R&D. These rules were already in place for R&D spending, and were generally welcomed by companies with actual R&D expenditures.
The problem with section 174 is that it essentially forces ALL expenditure on software development to be treated like R&D.
Being formerly in the materials industry at an individual contributor level, I think Asia’s preponderance of materials development firms is less about what the U.S. incentivizes and more about how poor the quality of life is for the engineers compared to what they could be doing. Asian engineers will put up with a lot more B.S. than Americans, and they have low birth rates to show for it.
Since this is already hanging around for an entire tax year a lot of companies are 1/5th into that pain. If this does not get changed in the next few months it would not surprise me if this becomes the new norm. Which would also be quite interesting to see how that would play out. Some companies apparently have already been amortizing salaries for a while in anticipation of this (eg: Google). Given that this also greatly punishes outsourcing I would not be surprised if at least that aspect will remain even if some of the rest will be rolled back.
> Some companies apparently have already been amortizing salaries for a while in anticipation of this (eg: Google)
> Given that this also greatly punishes outsourcing
Anyone help me understand these more? My understanding was that instead of deducting the costs of paying software devs the year it happened, it will be spread over 5 years. Which leads to a bigger tax bill now, and benefits bigger companies with deeper pockets as opposed to smaller businesses which have to raise moneny to pay taxes (or lower costs, potentially lower hiring). This should also push companies towards outsourcing, since not all places have similar laws? Is my understanding wrong?
This applies to all foreign R&D, not just where it is much cheaper… and unless it is many times less expensive, it will not be worth it. Only being able to deduct 1/15th per year is absurd.
If you're not yet profitable it doesn't matter, so you can do this accounting and once you are profitable (in five years) you'll be caught up
What the (possibly temporary) temporary law probably actually does is decrease the incentive to become profitable for the next couple of years (assuming your business is strong and you can raise another round)
Is it true that they are 1/5th into the pain? What about every new hire? That's the part I don't understand. It seems to discourage companies from increasing their headcount. Also, what happens when an employee leaves after 2 years? The company paid 2 years of salary but expensed only 35% of year 1 and 15% of year 2.
Update: Now thinking about it, it doesn't matter if an employee leaves, since the company will expense their salary portions that they haven't expensed yet in their future tax bills.
You already have such amortizations for a lot of things. In some cases this even gives you possibilities to improve your tax burden. It just means that you cannot deduct it all in one year. If you downsize a company to zero employees you still get to subtract salaries for a few more years against your profits.
It will set different incentives wiring wise and I’m not convinced they are good ones, but from this rule some people will benefit so they might fight the rollback.
Would this cause more intentional hiring? In other words, since everyone's already 1/5th into it this means that anyone who was hired in the last year since all those layoffs was hired intentionally with more of an expectation of betting on them for the longer term.
>A handful of these small business owners have bravely spoken to journalists from the Wall Street Journal and CNBC[...]But the vast majority are hesitant to speak to journalists as it might give their competitors free intel or make their employees nervous that they might lose their jobs.
This seems a tad overblown? "brave", "free intel"? Why would we need to hear the same thing from thousands of small businesses? This stuff is all over the industry, but employees are oblivious unless their company is explicitly interviewed?
It's a bad law, but I'm not sure what is interesting about this Tweet.
As an employee, if I hear my CEO/CFO saying how this law makes it hard to employ SWEs, I will be likely to worry about my job.
I pay attention to news coverage and interviews of my company execs, whatever they’re talking about. I don’t pay attention to all other company execs talking about tax law.
> Why would we need to hear the same thing from thousands of small businesses
That's how lawmakers prioritize work.
It seems some kind of lobbying organization for early stage companies might be needed - maybe some think tank funded by a couple early stage focused VCs
If this persists, it seems like it’d make sense for many software companies to shift the development and ownership of their software outside of the US, and then pay licensing fees to those overseas subsidiaries for the use of the software, which isn’t taxed as R&D.
Why on earth would someone give up ownership of their product or platform to an overseas subsidiary? The risks of that are massive, especially in a setting that is more prone to conflict as of late. This would increase the risks of losing your IP entirely. If ownership and development move overseas, then you'd be giving what you have away to someone else to entrust them with the entire thing to avoid paying a bit of tax. Software has its own supply chain too, and if you lose your product or platform due to doing this then that's going to be too bad...USA protections means USA presence and USA taxes.
TL;DR: Engaging with decentralized organizations and open-source communities allows access to global talent and diverse perspectives, fostering innovation and resilience. This approach, focusing on collaboration over ownership, offers potential benefits in R&D and innovation, outweighing traditional models that prioritize geographical and IP constraints.
1. Global Talent and Diverse Perspectives: DAOs and open-source projects, by their nature, often operate without geographical boundaries. This allows them to hire and collaborate with talent from all over the world, bringing in diverse perspectives and expertise that can be crucial for R&D and innovation. The traditional model of keeping all operations within a single country might limit access to this global talent pool.
2. Decentralization as a Strength: Decentralized structures can offer resilience and flexibility. In a world where geopolitical tensions and conflicts can disrupt traditional business operations, a decentralized model, with no single point of failure or control, might actually reduce certain risks. Intellectual property, in this case, isn't concentrated in one jurisdiction but is part of a global network, which could mitigate the risk of loss due to regional conflicts or regulatory changes.
3. Innovation and Experimentation: The open-source and DAO model is fundamentally about experimentation and pushing the boundaries of what's possible in technology and organizational structures. By embracing these models, companies can participate in cutting-edge developments and explore new ways of working that might not be possible within the confines of traditional corporate structures.
4. Intellectual Property Considerations: While there are legitimate concerns about IP protection, decentralized and open-source models often operate on a different paradigm regarding IP. The focus is less on ownership and more on collaboration, community, and building upon shared knowledge. In many cases, the value generated isn't from the IP itself but from the community and ecosystem that develops around it.
5. Regulatory and Tax Implications: It's important to acknowledge that regulatory and tax environments are significant considerations. However, for some organizations, the benefits of global collaboration and access to decentralized structures might outweigh the simplicity of operating within a single jurisdiction.
I'm a fan of DAOs for some things, but giving the legal ownership of the IP and the keys to running the entire platform to a subsidiary is not the same conversation as "here's how DAOs work."
This is exactly why I decided to not make my single-member, zero employee LLC an S Corp; if I'm paying no salaries, I don't have the Section 174 craziness.
With an S Corp, you have to pay yourself a salary, and if your work is primarily software development, Section 174 might hit you. (IANAL or accountant. Talk to one of them.)
Unfortunately, this doesn’t erase the impact. This applies beyond salaries to all resources used for what 174 considers R&D. Servers, software, the desk chair you sit on to do development —- all have to be amortized under 174.
It also applies to e.g. saas used for software development, which never would be depreciated over time previously. So your paid GitHub account? Amortize it over 5 years.
This change was actually part of the Trump tax cuts in 2017. It was delayed by five years as an accounting trick to make the bill look better at the time: the tax cuts’ projected long-term impact on the deficit didn’t look quite so bad when they tacked on a bunch of tax increases that would take effect in the distant future of 2022 (and with the assumption that this could of course be repealed if Republicans stayed in power).
Why target software companies? I guess because their owners look more like Democratic voters than, say, real estate investors who benefit from massive tax breaks that remain untouchable.
Yeah, the "Coastal liberal elite" is a popular boogeyman. I mean, look at the actual "elites", don't see too many Bernie-bros alongside Musk, Thiel, and the Zuck.
The same continuing income occurs for journalists and other fields that produce copyrighted output, in fact it's more true there since there isn't a substantial bug fixing load.
So why just software and not all fields that produce durable intellectual property?
Also it made me think about the converse, which might be interesting - libre software. Assuming a company has no plans of dual-licensing, ownership of the actual copyright of libre software is independent of how that software gets used to generate income. Anyone else could come along and use a copy of the software to create a business without any license payments to the owner. Or alternatively the original company could donate said software to a nonprofit steward (eg FSF or Apache) and be in a similar position. So accounting wise, it would seem that amounts spent on developing libre software could be more appropriately classified along with things like recruiting, advertising expenses, or even charitable donations, rather than the creation of an income-producing asset.
Setting up that software to work on the company's production infrastructure would still be a development expense requiring amortization. But the work to keep it running would still be maintenance.
Yes Section 174 is a huge issue for US companies of all sizes. We are seeing a shift in companies and other decentralized projects explore places like Panamá. The blockchain industry is dealing with IRS, SEC and other regulatory challenges. Some companies are choosing to leave the US.
There are options and advantages to operating outside the US. US timezone, a USD economy and a territorial tax system makes it easy to operate from my home country.
The IRS provided incrementally helpful guidance on the types of costs that must be capitalized in September. The capitalization requirements are generally more restrictive than GAAP where only ~30% of costs are capitalized for book purposes at many companies. There are some types of software development that are outside the scope of 174 though such as UI changes that don't add new capabilities.
This makes no sense and impacts a lot of early stage software projects.
In many cases for bootstrap ones it makes them not financially viable - 100%+ effective tax rate would do that if you optimize for early profitability.
Even for VC backed ones, those taxes just eat into your runway. I know a bunch of companies that opt for some weird international setup to try and avoid the effects but it’s really not what you should do in early days…
No, what you linked to there specifically says that "In the case of a taxpayer’s specified research or experimental expenditures for any taxable year", _then_ any software development is classified as R&D for the purpose of requiring capitalization.
I don't see anywhere saying that _any and all software development_ is considered "research or experimental". Ie., if some company pays you to build a piece of software and you hand it over when you're done, you were not engaged in research/experimental expenditures if you paid people to build said software.
> Ie., if some company pays you to build a piece of software and you hand it over when you're done, you were not engaged in research/experimental expenditures if you paid people to build said software.
Then it's your customer's R&D, and they have to amortize their payment to you. If you're doing outsourced software development for SMBs and startups, you might not be directly taxed under this rule, but your customers are, which significantly affects the market for your services.
Absolutely true! It's a burden on our clients for sure, and something I'll be sure to make them aware of before signing on (presuming this section doesn't get unturned in the near future...)
> For purposes of this section, any amount paid or incurred in connection with the development of any software shall be treated as a research or experimental expenditure.
I've been researching this for another issue. Anyhow, expensing R&D is a perennial political issue. Logically, if you're creating an income-producing asset, then you have to depreciate it over its useful life. The theory is sound, but of course politicians can mess with it, as they do.
But I thought the "useful life" of software was 3 years. No?
You usually want to incentivize R&D, messing with that simple theory is the standard.
Anyway, even that 3 years common knowledge number is biased into long-lived code. Do you spend work time putting out fires? Do you write code that is immediately discovered to be wrong so you have to take more time and fix it? Those 3 years are only for software that already passed through all of that.
(Anyway, I don't have any interest on that fight, I'm just watching from a safe distance.)
> Does "maintenance" on an amortized asset have to be amortized as well?
Usually not. Also, from other comments here, it looks like bug-fixing doesn't count as R&D (anyway, in a development focused company, you'll probably spend more deciding what is or isn't bug-fixing than you'd get from the difference).
Still, there are many more things that make software disappear just after being written. It's actually similar to most of R&D, so it makes little sense to count it the same way as capital investment.
The real kicker to this entire discussion is this: In a time where America is ostensibly gearing up for peer competition with China and working to onshore manufacturing and development, we are strongly disincentivizing domestic innovation by aggressively taxing R&D while China strongly incentivizes innovation with a 150% bonus deduction for R&D expenses[1].
Well the “good” thing about this law is that it also classifies ALL software development (and any supporting activities!) as R&D. And yes that’s as absurd as it sounds, but it’s true.
Your consultants were helping you get a tax credit for R&D under section 163. That has not changed - not all software development is R&D for section 163 purposes, and you still need to justify why you're classifying software development as R&D for section 163.
What has changed is section 174 - previously all software development could be expensed (without any justification needed) and now it must be capitalized.
My understanding[0] is that building product was always categorizable as R&D (though it was up to the company to choose to categorize it that way). What was not R&D was operational expenses (production support, etc).
[0] This is based on working at a company that went through a brief period where we had to do extra tracking to be able to capitalize our time if it was possible, and then reading up on the matter to try and understand what the hell the purpose was. Possible I'm wrong about the matter--I've never been the one doing the finances.
This is actually great if true! Maybe companies will finally be incentivized to start fixing their bugs rather than cramming unwanted features and UI re-designs into their software products!
However I doubt they'd leave such an obvious loophole. Surely the IRS's definition of R&D includes all types of software development activity.
It's not an "obvious loophole". They are quite specific about this. New features, new capabilities: R&D, must be amortized. Bug fixes for existing features: expensable.
The gray zone comes when a bug fix actually provides a new feature (e.g. something that wasn't actually usable before).
This explains why I have seen so many tech companies doing layoffs. They layoff now, wait for the fix(hopefully), then hire people again(hopefully). And here I thought it was AI replacing the workers.
Is anyone here capable of cogently summarizing The Tax Relief for American Families and Workers Act of 2024, its likelihood of passing into law, and whether that will actually fix this?
(As far as I can read it, if passed it would delay this until 2025, likely with the intent to change the laws before that but with more time to debate.)
How typical of Congress to intentionally set up fights every 2-3 years over completely obvious BS, and summarily shut down the government and scapegoat something else like top-line rates on the "working rich" while playing class warfare politics.
We're all going to end up working for dev agencies that build stuff for IP holding companies in Panama, which then license the use of that software back to their parent American tech companies.
We need temporary relief, immediately. But long term, how about we shrink the size of the federal government? Then we don’t have to flog small businesses to death to feed the massively inefficient and wasteful machine!?
These changes came about as part of the Trump administration & Republican efforts to reduce taxation. Because you can't trivially just cut taxes (as was done in 2017) without showing how you will either make up the missing revenue or reduce spending, they threw in a time-deferred measure - the new Section 174 that "demonstrated" how the tax cuts in 2017 would be paid for in the future. It was all sleight of hand, but what's new?
Republicans have been talking about shrinking the size of the federal government for decades. They routinely fail to come up with substantive cuts in spending that would not set off a disaster for them at the polls, or for the country economically, or both.
There is this perception that the federal is full of wasteful spending, but this perception is false (there is waste, it's not a large percentage of the huge overall federal budget). This means that to actually cut the revenue needed will require truly substantive cuts to the services provided and performed by the federal government.
Maybe you could be more successful at proposing what those cuts would be than any Republican has for 50 years, but it seems unlikely.
US voters have consistently rejected such policies for nearly a century, if not longer.
While there are certainly people who believe as you do, you are not in a majority. There are also a lot of people in Argentina (possibly a majority) who don't agree with those policies either.
IANAL. IANAA. As a sole proprietorship it's likely very good for you (until the tax code is fixed). Your potential clients can write off your subscription costs as immediate expenses, rather than building in-house and amortizing the R&D costs over 5-15 years.
As a different entity type, if you pay yourself a salary, I believe you amortize that like any other company would. But if the U.S. is anything like Canada, you can pay yourself a much lower salary, and then pay yourself through dividends as well.
This is a much better article: basically these companies are paying for the tax cut Republicans gave the to the top bracket in 2017:
> In 2017, then-President, Donald Trump, signed the 2017 Tax Cuts & Jobs act, which overhauled tax codes and reduced tax – for example, it reduced the top tax bracket from 39.6% to 37%. To make the bill pass strict budgetary rules, the Senate used a process called reconciliation: adding in tax code changes that delayed tax increases. These delayed increases “balanced out” the tax reduction.
This is correct. Since it was basically accounting sleight-of-hand to get a better score from the Congressional Budget Office, they knew this would be disastrous if it took effect, and were widely expected to repeal it before it did. That deal, which also paired fixing it with the child tax credit, fell apart in December 2022.
I think of that every time someone proclaims that some time bomb like that will force the Congress to make a deal. That’s the instinct of a bygone era and I no longer feel safe counting on it.
Yeah… that’s why this is still far from a guarantee, even though it’s a very positive development. This could very well still fall apart like last time.
Everyone, call your Reps and ask them to tell leadership to support the tax deal!
You're getting down voted for sounding partisan I suspect, but regardless of the politics, you're absolutely right. That bill was a huge change across the US tax system.
Yeah, it’s an election year so even a literal statement of fact is going to be downvoted by some people. Hopefully at some point they’ll have an epiphany about whether they should self-identify with a party so strongly when an accurate descriptions of its actions feels like an attack, but I’m not as optimistic about that as I used to be.
It’s a partisan statement since you only mention the upper tax bracket and not that it also slashed taxes for the lowest one completely and reduced them for basically everyone while making changes to a ton of other things (like child tax credits) as well.
You’re over simplifying and singling out your pet political issue in an effort to sway opinion to your political leaning. If you had just said it was due to those tax cuts you likely wouldn’t have been downvoted.
Most of the need for the reconciliation was the cut to the top tax bracket, where most of the the tax revenue comes from (e.g., the top 1% of earners pay 42% of the taxes [0]). So focusing on that bracket seems valid in this context.
The reason the author I quoted said that is that most of the cost comes from the top. If we were taking on a couple trillion in extra debt, that seems like a rather minimal benefit to give ⅔ of it to the richest people.
Why would I be on the side of someone willing to go through all this trouble just to avoid paying American wages? Frankly, I would prefer you not start a business or do it elsewhere.
The problem with taxing software development like this is it is impossible to know the value of the asset up front. A $1M investment might be worth $1B or $0.
is extremely clear on this. New features, improvements? All costs must be amortized. You can only expense bug fixes and GUI changes that don't add features.
Let's just say I started a company last year. I sell software and made $100k revenue this year. I hired someone for $100k to help build that software. I have zero dollars left in the bank at the end of the year.
According to 174 I can only amortize that $100k over 5 years, not all at once. So now I have to pay taxes on $80k profit. I have no money in my bank account, how can I pay any taxes at all?
It's not like I'm going to fire my engineer, he does feature and maintenance work. Let's say that next year my business stays the same. 100k in 100k out. I still have no money in the bank and have to pay taxes on $80k profit, again?
What??
Every little startup in Silicon Valley makes software, spends a lot on salaries, and is barely breakeven. How... am I expected to run a business like this?
Not to go all achstuushually on you, but it's even worse.
You can only deduct 10% the first year. Every year after is 20%, then year 6 is a "make up" of the last 10%.
So year 1 would be $90k now not $80k.
> Because of the tax code’s accounting conventions, domestic firms ultimately have to deduct 10 percent of costs in year one, 20 percent of costs each year in years two through five, and the remaining 10 percent of costs in year six.
You're understanding it correctly. The "answer" is that you either take VC investment, borrow money to pay the taxes (in this economy??), or go out of business.
What? Why is all software work counted this way? Why software? Shouldn't I be paying taxes later once I've made money? This is like paying taxes on unrealized gains - I can't pay those taxes - I haven't made any gains!
Because it was used to game the CBO score for the 2017 tax bill by offsetting tax cuts with future revenue, making the bill seem more revenue neutral upon analysis. This was done under the presumption that Congress would never actually let something so insane and detrimental to American innovation and competitiveness go into effect, and would obviously roll it back before 2022. The IRS didn’t even have guidance for the change until months and months after the April filing deadline, because why would they prepare guidance for something that will so obviously be rolled back?
Then Congress did not roll it back. So here we are.
Wait, isn't it an incomplete example? How much money did you start with? In other words, you had to pay the developer some salary before any revenue came in.
I built a prototype for a potential customer and they agreed to pay me 100k to develop it. Let's say they're my only customer - we're a boutique shop. I didn't need any startup capital - just needed to promise to solve one person's problem.
That comes off the balance sheet which doesn’t impact taxes. Eg if they raised $100k from friends/family to start the company, that equity investment has no bearing on the P&L or taxes owed.
You have to have income in order to be taxed. So there is no taxing non-existent profit. There has to be actual profit. What people must be complaining about is that they can't take all the income from product X in year Y and say it was spent on developing product X so no profit was made. Instead they'd need to have spent that income on marketing supporting and maintaining that product. Then they'd be making zero profit and no tax owed.
If you paid 100k in salaries and earned 100k in revenue, then you broke even, and no actual profit was made. Under the rules being discussed here, however, the 100k you spent in salaries has to be amortized over five years, meaning that even though you actually spent 100k this year, only 20k is deductible from this year's taxes. The IRS then taxes you on the 80k of "profit" you made, which doesn't actually exist, so you don't have the funds to pay the taxes, and your business becomes insolvent.
This is absolutely about taxes being charged against non-existent profit.
That “income” is being legislated into existence by disallowing the deduction of customary primary operating costs that are allowed in any other business.
In effect, it becomes a tax on revenue in the early years of a software company. Taxes tied to revenue are notorious for creating unhealthy and perverse market incentives, which is why very few jurisdictions use them. Many small software businesses are now in the position of potentially literally paying more income tax to the IRS than the profit + asset value over the entire life of the business. It is pretty messed up to make small business owners pay “income” taxes out of their own pocket in the absence of offsetting profit.
It is yet another indefensible and unique perversion of reasonable policy under US tax authority.
The best solution is to incentivise investment by setting its tax rate to 0. That's what Trump's cash-flow tax would have done. Higher taxes on the money printing ventures (Google ads etc.), lower taxes on start-ups and companies developing new products. That would be besides all of the other benefits like preventing profit-shifting.
Maybe if Silicon Valley wasn't so intensely ideological it would have backed the right horse when it came to tax policy. I have very little sympathy here because I think most software developer salaries really are 'development', the problem isn't the classification of developer salaries, it's that the current corporation tax is dumb.
Yes and I'm talking about the original 2016 house-republican DBCFT proposal which would have eliminated tax on investment in general.
Section 174 is a hack to fix a moronic tax that would have been replaced if not for the left. There shouldn't be a section 174 to even amend. And as I said, software development clearly is development, the change makes perfect sense.
If you want R&D to be taxed (even with deductions which get amortized), and you work in R&D, you don't get to whinge when a loophole gets closed so you get taxed like other R&D expenses. The people getting bankrupted by this deserve it insofar as they opposed the tax reform bill that would have elegantly solved this problem forever.
Who does this benefit in a zero-sum competition scenario, assuming that everybody is affected.
For instance, imagine a competition over the AI toaster market. You have Philburn, the broke PhD student with a great idea, Burnright LLC, the toaster algorithm and electronics designer, BurnCo, the manufacture, and B2N, the toast technology venture capitalist.
A few scenarios: 1) BurnCo hires Philburn for internal R&D. 2) Burnright develops the tech and licenses to BurnCo. 3) BurnCo contracts Burnright for R&D. 4) Philburn starts Smoke, gets B2N investment, and gets acquired by BurnCo.
Which of these, or others, become comparatively advantaged with this new tax treatment?
> Who does this benefit in a zero-sum competition scenario, assuming that everybody is affected.
Foreign-run software companies not affected by this law, big companies with deep pockets that don't need to fear upstart competition, and other industries that are not R&E but can enjoy the lower tax rates from TCJA that were paid for with this revenue.
Playing devil's advocate here. There are a large number of jobs that fall under this R&D classification in the software space that I do not feel carry the weight that a term like R&D ought to imply. This carve out was clearly intended to be for fundamental research in hard technical disciplines. Much of where it is being claimed however is really more accurately described as product development and market research at best. Not the kind of hard technical R&D that this is pitched as.
It seems that the number of software engineers involved in actual R&D (according to my admittedly made up and subjective definition) is vanishingly small, much less than 1% in all likelihood if we do some quick napkin guestimates.
So to what extent do we want to provide tax incentives to (an already profitable in some cases but also commonly accused of setting big piles of money on fire) industry for normal run of the mill product development? That's what's really going on here imo, and it's not obvious to me that there is any way to justify that. There has to be some line drawn somewhere between applications development / product development, and true R&D ala some systems research group working on scheduling algorithms or whatever.
I admit that the line here is blurry and not well defined. That needs to change if there are tax incentives involved.
At the very least, anything that you know for sure is going to end up in an end product is NOT R&D. The definition ought to depend on what a reasonable and informed observer would classify the risk of total failure to be. True R&D tends to have a very high risk of complete failure in either a technical sense or a product integration sense (as in, you find a solution to your question but it cannot be made into a commercially viable product, often for technical reasons or the specifics of your solution). That is what laws like this are intended to incentive, because we have collectively decided that we want this research to take place even if the risk calculus is such that it is entirely unprofitable for a lot of companies.
I have said this before and attracted downvotes for it, but here goes.
There are two possible arguments here to try, and I have only ever seen people lobbying against this change use the one, less persuasive argument. That argument is what I will call the "incentives" argument: that change in the rule provides bad incentives against doing R&D. This argument goes along the lines of "this will cost jobs for R&D workers" or "this will reduce the competitiveness of the US." The other possible argument (that I have not seen cited) is that "R&D" work can be operational, and that forcing capitalization of R&D expenses is a bad accounting practice. This Twitter thread only argues the former.
The glaring problem with the incentives argument is that it gives up the point that the ability to operationalize R&D work is a subsidy for technology and software companies. This is equivalent to asking for a subsidy at a time when startups can pull $100 million with no product and other technology companies are reaping record profits. That is not a particularly persuasive argument, and it triggers bad emotional reactions from people. If not for the SBIR companies getting absolutely shafted, the responses to this argument I have seen from non-tech people range from "fuck you" to "deal with it."
The accounting argument is boring and sort of technical, but also a lot harder to argue against and doesn't trigger a negative emotional reaction. All of the people making the rules will understand it, and the IRS could even make the clarification on what is and isn't "R&D" on an accounting basis without an act of congress. They have kind of done this, but have not been pushed nearly far enough.
Front loading taxes on a new business makes little sense no matter the industry. Long term view is that more taxes are generated from a surviving, healthy business.
I (shamefully) have an MBA, and I learned a lot about accounting during that process. Whether basic research ought to be operationalized or capitalized when looking at company valuation was actually a bit of a point of debate, but both sides have strong arguments. What definitely shouldn't be capitalized is "new feature for my existing software" which pretty clearly is capitalized under the current rules.
For one, companies will be encouraged to classify R&D as operational expenses where they can. This distorts the overall picture of the business’s fundamentals.
It's probably more than that. You have external GAAP, non-GGAP, constant currency vs. non-constant currency metrics, all sorts of internal accounting measures. With larger companies it's complicated. (Smaller companies it's more about cash flow even if they consider taxes as well.)
>the responses to this argument I have seen from non-tech people range from "fuck you" to "deal with it."
Two factors also don't help matters either:
* Most people, which perhaps surprisingly includes most commenters here, don't understand what taxes are levied against or what losses and amortization even are. You can't effectively argue against something you don't understand.
* Of the people who do understand taxes and income/loss sheets, it's only the business owners and accountants who like putting as big a number as possible under losses to reduce net profit and thus taxes owed. The commons rightfully see it as tax avoidance, which unsurprisingly is met with "fuck you".
"tax avoidance" Otherwise known as what every sensible and smart person does because that's not the "governments money".
AKA: "How dare people want to keep their own money and not freely give up larger and larger percentages to ever larger government programs that are inefficient cesspools of mismanagement and corruption".
I'll match people saying "fuck you for wanting to keep your own money" with a fuck you for supporting the destruction of businesses on the notion that it's the governments money in the first place.
Can't you only deduct the mortgage interest? Or are you referring to something else? The mortgage interest deduction turns out to not be very much in practice, in my experience, anyway.
It used to be a bigger deal. Yes, housing prices (in some US geos) have gone up a lot but so have standard deductions and interest rates are still relatively low historically. Absent relatively large charitable donations, big mortgage interest payments, etc., itemizing doesn't make a lot of sense for many people.
Your comment is completely backwards. Chevron has nothing to do with this.
This was a specific law passed by Congress and the Trump administration to make their massive tax cut for billionaires revenue neutral.
The IRS is implementing the law as passed and has nothing to do with Chevron. It would be the exact same with or without Chevron because this has nothing to do with it.
It’s quite telling that the anti Chevron people have to rely on blatant falsehoods to push their agenda.
Am I correct that if this does not get fixed, it's a huge boon for dev agencies as paying them is an immediate expense for the startup, while paying in-house devs is not?
From my reading, that would be an expense directly related to the development of the software, which also must be amortized.
What I haven't figured out is what the Dev agency has to do. They don't own the software being developed. Unless they have in-house libraries they develop...
And how does this work for open source projects that have corporate support to be developed on during company time? If you amortized the "R&D excited" of the developer's salary on the premise that the software will make long-term, recurring income, what does that mean for software that isn't being sold?
My previous job, I was the only software developer, making a tool that employees and customers in the company used together. We didn't sell the software, we sold a service completely unrelated to the software, the software was just a supplement to the service.
Sorry, Im pretty anti-tax, but this is the mature way to run a software business... I remember starting my career and presenting amortization sheets for software R&D and not a single exec giving a shit, whole industry gave up if you ask me.
Please explain then. Because if you spend $2m on an idea, layoff half the company, pivot to a new idea, rehire. You think it's fair on the rest of us you get to write that risk off? This. Is. Why. People. Hate. Tech.
Maybe? What if it was a one time manual test harness? What if it was a migration for a new client that has since quit? There is a pretty wide range of tasks a developer can do.
With a startup you're running a search function. The code you hired the developer to write often does stop working because you've been forced to pivot as you search for product-market fit.
If we spend $1M in year one to build a platform, we will surely spend more than $1M in year two maintaining and extending it, and keep doing that until almost the day we finally abandon it. Depreciation might make sense when the actual investment was heavily front-loaded (compared to the revenue) but in our industry it almost never is.
Pretty much everyone else works on the idea that salaries are operational expenses and not subject to amortization.
I have yet to see any other case of salaries being turned into Capital Expenses (aka something that requires amortization/depreciation). I'd understand if outsourcing (or otherwise contract work) would be treated as something to be amortized, but we're talking SALARIES.
I don’t see why software should receive special treatment when 99.99% of software engineers/developers are not in any way engaged in what might be considered genuine scientific research and development, which R&D tax advantages should be reserved for.
> Changes to R&D amortization were a rude surprise to them, as they'd never had to amortize software development before, and didn't think of the work they do as R&D.
Precisely. They didn’t think of their work as R&D because it was not R&D. Frankly, they should have seen this coming.
> These are small businesses we're talking about. Almost all of them make under $10M in annual revenue, and the vast majority are under $2M in revenue.
What about the millions of other small businesses that don’t get special tax treatment? Restaurants, bars, plumbing companies, landscaping companies, accounting firms, etc.? Software engineers are not scientists. At the end of the day, a company is supposed to be able to stand on its own two feet, not rely on government handouts. Allowing otherwise unprofitable businesses to stay in business disincentivizes innovation and efficiency, which harms productivity growth and makes us all poorer in the long run.
>Precisely. They didn’t think of their work as R&D because it was not R&D. Frankly, they should have seen this coming
Don't you have this backwards, or do I?
Software development is now going to be treated as R&D, which means that costs have to be amortized. That's not an tax advantage, it's a disadvantage. All the other businesses you list get to expense their labour costs. But we agree, software dev should not be treated differently.
MasterYoda900 has it backwards. However, my CPA said software engineer salaries are only categorized as R&D before the product is launched. So this only affects new startups pre-launch... if my CPA is correct.
Your CPA might be referring to rules for R&D tax credits, which covers a much, much more narrow scope of business activities. R&D credits cannot offset the impact of the amortization rules.
> However, my CPA said software engineer salaries are only categorized as R&D before the product is launched
I wonder how this one will pan out.
Let's say you build out a landing page with a way for folks to input their email address which demonstrates interest in your product. You've released it and people can sign up.
Did it launch? You could make case it did and now you're iterating on the product from here on out. You've launched phase 1 of the product which is probing for demand.
Any software development costs related to improving the product by providing new features must be amortized. Only clear bug fixes can have costs that are expensable.
This is made extremely clear in the IRS guidance document.
They have it backwards in two ways. Software isn’t Research, but Software Developers are generally doing Development ie improving existing products or developing new ones.
I used to have similar mindset about this like you, but let's read what's an actual definition of R&D:
>Research and development is the set of innovative activities undertaken by corporations or governments in developing new services or products, and improving existing ones.
>R&D activities differ from institution to institution, with two primary models of an R&D department either staffed by engineers and tasked with directly developing new products, or staffed with industrial scientists and tasked with applied research in scientific or technological fields, which may facilitate future product development. R&D differs from the vast majority of corporate activities in that it is not intended to yield immediate profit, and generally carries greater risk and an uncertain return on investment.
and as you see "R&D department either staffed by engineers and tasked with directly developing new products, or staffed with industrial scientists and tasked with applied research in scientific or technological fields, which may facilitate future product development"
Note that the current tax code implies that all software development requires amoritzation of salaries. This is as far as I can tell internationally quite unique. Usually you can write off all salaries as expenses.
I don't have a fully formed opinion on this aspect of the taxation, but it's pretty evident that this is out of the norm.
>software engineers/developers are not in any way engaged in what might be considered genuine scientific research and development
Have you forgotten the D in R&D? Software engineers are building the product. That’s development.
What do you mean by “genuine scientific research”? I suppose you think ML researchers aren’t doing “genuine science” because they wear hoodies instead of lab coats?
Yes, if your company has builders that build you an asset (like a warehouse or a factory or office building) then you always had to amortise those salaries together with all the other expenses for building that capital asset.
That's not the case if you contract your builders out to build a house for someone else and they pay you for the hours, but IMHO it's also not the case (even with the new changes) for software developers being 'rented' as contractors to build stuff for others and the company getting paid for their hours.
I think the theory is that you can't keep selling the same house each year so you've just developed inventory and not long term assets.
Though, the country where I work and live, R&D has more profitable tax treatment, your employee costs are a same year expense, but you can get some relief on the rest if your tax bill if you proactively document what's R&D about your work and it doesn't get rejected by the tax department
Do your builder developers build the same thing over and over again like builders who build identical houses over?
I get the point but software really is different. Rarely ever are you building the same thing over, both from a “structural” or “architectural” perspective and from a functional perspective.
If your builders are creating a brand new blueprint for every house /structure and designing everything from scratch then I think it’s more similar, probably closer to commercial property builders with large structures.
Now under that perspective, I still don’t exactly understand why you’d treat labor cost deductions differently. Maybe the fact that these more complex products and services being built have higher return potential? I don’t know, I’m searching for a rational reason.
But, I do disagree treating software like something as commodified as many forms of labor. Someday we may get to that point with simple principles and procedures to follow, albeit technically, but we’re not there yet. At higher levels that may not be much “research” that’s needed but where the rubber meets the road, your engineers are checking to see what exists and doesn’t, designing new solutions to fill gaps in between, testing things, digging into previous things that don’t work and trying to fix them… it’s research.
As we continue to move more professions to an intellectual economy, many roles are becoming more and more like this, not just software. I’d say loading more roles and expectations on labor is also creating this situation where we’re asking more people to do more things they don’t know or can’t possibly know everything at a proficient level and the things we’re asking them to do don’t have solutions that can be written down and quickly referenced when needed as a simple recipe to follow. If you make the recipe general enough (e.g if a problem arises, find a solution and fix it at a low cost) then one might point to such highly generalized solutions and pretend it’s a solved problem but it’s not, it lacks any useful concretization. “You're not doing research, I told you ‘if a problem arises fix it’! There’s nothing to research you just do that!” doesn’t cut the muster.
I don’t think you understand what they’re discussing.
If you run a plumbing company, make $100K in revenue, pay a plumber $100K, you have no profits, owe no corporate taxes.
If you run a software company, make $100K in revenue, pay a software engineer $100K, section 174, which just went into effect recently, means you now owe taxes on $90K of profit. This is because you must spread dev costs over 5 years, you can’t deduct them in the year they happened.
The code defines that you start with the midpoint of the taxable year, so even though it covers a five year period, it will also cover 6 tax years. Because of that, the first tax year (which was 2022), only 10% can be deducted. Tax years 2 through 5 is 20%. You can deduct the last 10% in the 6th tax year.
You only get taxed on the revenue. So no, if you don't have revenue, you're not paying taxes. But all those businesses CAN deduct their expenses, and only pay taxes on profit.
Software is now in the unique position of being required to pay taxes on revenue instead of profit. If you spent $5M on development and raked in $100k, you'll be taxed on $100k (minus expenses, where the $5M must now be spread over 5 years). Of course, with those numbers, the company might want to spread costs over 5 years anyway (which they could already do).
So really, it's as if the tax code is specifically targeting bootstrapped companies that are reinvesting their profits back into the company. IMHO, the worst possible option of any that could have been chosen to milk software companies. Large established companies can afford it, VC companies were going to amortize anyway.
Yeah, you "eventually" can expense your costs. But that just means that it's a tax grab on the smaller companies that will be put out of business by this move. They pay all the tax this year, go out of business, and never get to reap the profit of the development work.
> I don’t see why software should receive special treatment when 99.99% of software engineers/developers are not in any way engaged in what might be considered genuine scientific research and development, which R&D tax advantages should be reserved for.
Why? There's a clear relationship between R&D spend and economic growth. This is including R&D spend on non-scientific endeavours like commercial products + processes, which the majority of non-maintenance software work falls into.
Estimates vary, but I've never seen a cost/benefit analysis of unrestricted R&D that finds anything lower than a $2 return for every $1 spent in unrestricted R&D.
