I can understand how a sudden unexpected change to the tax code could catch people off guard and cause short term problems but overall I don't see why this particular change should be so devastating once any transient effects have been absorbed.
I can understand how a sudden unexpected change to the tax code could catch people off guard and cause short term problems but overall I don't see why this particular change should be so devastating once any transient effects have been absorbed.
It is especially problematic since it categorizes all software development as R&D even if we don't think of it as R&D. It's still unclear what the IRS considers "software development" since they've never had to define it, but the way most big companies with their well-paid accountants are proceeding are that it covers new product development AND new features on existing products, but not bug fixes/maintenance.
Let's take a simple example. Imagine a profitable small software company that made $1M in revenue last year, spent $700,000 on developer salaries and $200,000 on other expenses. Ordinarily, they'd be able to write off $900,000 and have a taxable net income of $100,000 that matches their actual profit. Assuming a tax rate of 25% that's a $25,000 tax bill.
Now, if you assume developers spent 50% of their time building new products and new features, and 50% of other expenses were on new features, only $420,000 of the salary costs and $110,000 of other expenses are write-offs. Their taxable income just went from $100,000 to $470,000.
Assuming a 25% tax rate, their tax bill is now $117,500 for 2022 — which exceeds their actual net income. This also inflates their quarterly tax payments for 2023, both of which hit right now.
This gets even worse for companies that aren't profitable, as they don't have the cash flow to cover a tax bill when they hadn't planned on having one at all. And given the current financial environment, it's hard for startups to get any kind of additional financing or funding.
This news article about our effort gets into this a bit more: https://technical.ly/civic-news/section-174-small-software-c...
If this is really the way it works, defining some salaries as necessarily not being deductible from revenues, then it makes no sense for multiple reasons.
First the developers are still paying income tax on their salaries so that money is getting doubly taxed in the year the revenues are received.
Second the government generally seeks to encourage employment. This would have the exact opposite effect because any employee you hire who's doing software development would cost you (1 + 4/5) times their salary in the near term.
I wonder how much of the downturn in tech employment this year is being caused by this.
How can that be true? You only pay them once, not 1 4/5 times.
Not exactly. It's a well-established accounting principle that you capitalize costs that provide a benefit over multiple years. Depreciation is an easy-to-understand example. It's more true that the historic practice of expensing R&D costs was artificially inflating costs.
What the tax change is doing is forcing amortization, which, for early-stage companies is difficult, because they have depended on expensing early and recognizing income later.
It's a difficult issue. There are good arguments on both sides. But it sounds like this was a surprise, which is surely not optimal.
fwiw, when I was running start-ups (80s/90s/00s), my recollection is that we amortized our software development costs. I guess this got turned around by the rise of the sophisticated startup world, with more accountants, lawyers, and lobbyists. And now the government is pushing back, not without reason.
OK, so lets flip this. I'm a founder working for free, as many founders do. We code on nights and weekends and produce hundreds of thousands of dollars of capital value. If the business doesnt work out, can I claim all this as a loss?
We cant have it both ways, can we? So I should be able to take losses on these hundreds of git repos I have with thousands of hours of unpaid work?
I sold a business at a $60k loss 4-5 years ago and between capital gains offsets and normal income deductions I still have about $30k in deductions left.
There are a lot concepts being not very well defined here: employment, investing, taxation, salaries. It's not all one thing.
What is it that you think "we" are having both ways?
Paying yourself is a lose-lose game.
What reason is that? Increased tax revenue (in the short term at least)? Because if there's no difference in the long term then it seems pretty dumb to inflict financial turmoil for no net gain.
Is it? Seems like lawmakers just messed up in reaching an agreement to extend something that is usually extended. Typical congress games.
From light reading, Republican leadership seems to be the main blocker since extending the provision has bipartisan support. You would think that extending this and child tax credits would be no-brainers for Republican leadership, but here we are.
No, it is. That's literally why this happened, it was used as a bargaining chip/to buy time and they never cleaned it up. It's not supposition or a guess, it's the stated intent and consequence.
> it's not so obviously a good thing
Let's say you earn a million dollars before salary and you have 10 engineers working for you each making $100k. You pay out your salary and have $0 profit at the end of the year.
With this change, you are taxed as if you made $800k profit, so unless you've got a couple hundred grand in your bank account this is easily enough to bankrupt a business and put those 10 engineers out of work.
It would be one thing if the $800k was in the bank and this was Hollywood accounting to make it seem like it's not profit. But this is money that was paid to employees and now the business is expected to pay taxes on it as if it was never paid. It's absolutely farcical how anyone could look at this and not see it as ridiculous.
Unfortunately, I think a lot of hacker news posters give plausible deniability to that leadership in order to avoid cognitive dissonance with how certain political leaders that they support are not business friendly at all.
BTW, I founded and ran four companies and faced this very situation plenty of times. In my day (this was a while ago), we generally had to, or did, amortize development costs. I'm not saying I loved paying the tax bills, but the concept is neither farcical nor ridiculous.
Do you see a difference between software development in a consulting business model (instant one-off benefit) and software development in a saas product business model (benefit over multiple years)?
> There are good arguments on both sides. Can you provide the good arguments for capitalizing software development costs and not expensing it?
Can you explain the reasoning of charging taxes to a company that has revenue beyond merely 1/5th of its expenses (actually 1/10th in the first year, or 1/30th for international operations) and hence still heavily investing cash?
Yes. Not sure what that has to do with this discussion.
> Can you provide the good arguments for capitalizing software development costs and not expensing it?
Yes. The well-established accounting principle of matching income and expenses.
> Can you explain the reasoning of charging taxes to a company that has revenue beyond merely 1/5th of its expenses (actually 1/10th in the first year, or 1/30th for international operations) and hence still heavily investing cash?
Yes. See the answer to your second question. Companies often have to make investments. If they buy a Big Machine, they don't get to write it off in one year. There's nothing nefarious about amortizing costs over their useful life.
Because the law seems to be very strict that all software development should be capitalized. So would you suggest to split the revenue recognition depending on the corresponding business model of the product corresponding to the software development?
> If they buy a Big Machine, they don't get to write it off in one year. There's nothing nefarious about amortizing costs over their useful life.
It seems the capitalization of the wage of a software developer is being defended and put equal to the capitalisation of the cost of a Big Machine. I still see an unfair difference made in the reasoning. Let’s take following example
* Software developer has a wage cost in year 1 and builds a SaaS tool in year 1. The developer’s useful life w.r.t. the incurred cost is indeed 1 year and the revenue generating period of the product is 5 years.
* Big machine (crane) has a purchase cost in year 1 and builds a warehouse in year 1. The crane’s useful life w.r.t. the incurred cost is 5 years and the revenue generating period of the product is 30 years.
It’s being claimed that both the software developer's first year wage and the Big Machine purchase cost should be capitalised over 5 years. But that’s comparing apples with pears: either both should be capitalised over their own useful life w.r.t. the incurred cost (1 year vs. 5 years) or both should be capitalised over the revenue generating period of their product (5 year vs. 30 years).
Three other thought experiments:
* You should capitalize a crane when you buy it and you should expense when you (properly w.r.t. accounting principles) rent and use it for a year to build something. But when you rent a software developer (= hire) for a year to build something, that should be capitalized?
* When the crane is being sold or breaks down, you recognise a gain or loss and the capitalization stops. When the software developer leaves the company, is the capitalisation of the developer’s wage still continuing?
* I come work for you for the next five years as a software developer, but you have to pay me my wage immediately for the upcoming five years. Also you have to buy a GPU server that I will use to build my product and that will supposedly last 5 years. Are you capitalising both my cost and the GPU over 5 years? Or will you be capitalising my cost way longer than the GPU, even both I’m gone after 5 years and the GPU broke down?
The other option is they take a 1,000,000 loss that first year, and then pay tax on all $300,000 for each of the succeeding years. Either way, at the end of six years, There was $1,500,000 in revenue and $1,000,000 in expenses.
As far as the treatment of bug fixes, the rules around improvements and repairs probably cover that. If you fix a bug like a bad calculation - that's probably opex, like replacing a part on a machine. If you add a feature that extends the life of the product, like adding an API for outside developers, that would be an improvement and capitalized. This is like refurbishing equipment to extend its useful service life.
Isn't normal accounting principles usually that if a company pays $M salaries, then regardless of whether those salaries paid for an asset or not, they are an expense that's 100% deducted from the income when calculating taxes?
Are we saying that at a company with 2 desks where 1 is a marketing person or accountant and 1 is a software dev, their salaries would deduce differently from the company bottom line, because the software developer is said to create "assets"? Isn't the marketing of that asset likely to be build the value of it in the same way as the research and engineering does?
https://www.journalofaccountancy.com/issues/1999/may/maples....
IMO it does make sense to amortize software expenses like other capital expenditures.
If what you are doing is software development then obviously it is a development activity that falls within the meaning of development for purposes of tax laws.
Software programming that does not constitute development, such as bug fixing, is not subject to capitalization.
It's research (as in new knowledge) and development (as in new products based on existing research and knowledge). Software generally falls into the latter category. A scientific process is not required but does make it easier to document qualification for the R&D credit.
And yes, a carpenter developing new cabinet designs absolutely would qualify for the R&D credit (and their salary could fall under the scope of this rule change).
This is... Exactly what I said?
> And yes, a carpenter developing new cabinet designs absolutely would qualify for the R&D credit (and their salary could fall under the scope of this rule change).
I feel I pretty clearly alluded to the physical process of turning wood into a cabinet, not developing novel new techniques for doing such.
What makes it worse is that accountants at real deal firms like Plante Moran didn’t bother sounding the alarm early because they figured like every time in the last 70 years Congress would push off the effects.
It is an absolutely crushing situation that is going to put a lot of shops out of business unless they have cash on hand to weather the 5 year R&D tax amortization schedule.
Previously, $1.000.000 spent on R&D in 2023 would result in a $1.000.000 deduction on your 2023 taxes. Under the new system the same spending would result in a $200.000 deduction in 2023, $200.000 in 2024, $200.000 in 2025, $200.000 in 2026, and $200.000 in 2027.
You still get the same deduction, but spread out over multiple years. However, it also means that you can now deduct $800.000 less in 2023 than expected, resulting in a far higher tax bill this year! If you are a startup you probably don't have that spare $800.000 just lying around doing nothing.
It's bad that the tax treatment changes suddenly because businesses need stability, but the change itself is perfectly rational.
Profits = Revenue net Costs
Taxes are a cost. Taxes are defined as some rate t, tax = t * (Revenue net Deductible Expenses)
So Profits = Revenue - t * (Revenue - Deductible Expense) - Non-deductible Expense
Percent of t is small relative to the value of 100% applied to non-deductible expense. What this has done is to take salary, deployment infra, everything, from Deductible to Non-deductible expense, leaving 20% of what was there before. That is very large.
The issue is that the rules changed. Businesses that relied on the former rules are now faced with a (possibly insurmountable) challenge to accommodate the new rules.
Washington loves to fiddle with tax rules, and lobbyists spend a lot of time and money encouraging it, but nobody can anticipate the ripple effects. It all looks great on CBO spreadsheets and congressional press releases, but the real-world impacts can be devastating.
Note, this was not an "unexpected" change (it's been in the code), but it WAS unexpected that the provision was not extended.
Note that this affects not just startups. My wife's firm is a small, employee-owned, non-tech S-corp. This hit them as well. It resulted in tax bills for the shareholders approximately 25-30% greater than the firm's accountants expected them to be. The shareholders are on the hook for those higher taxes, although the company did the right thing and distributed extra cash to them to offset the higher taxes.
Now: deduct 20% of R&D salaries from income to calculate taxable profit, with the remaining 80% spread 1/5 per year over the next 4 years.
For software companies, where costs are basically eng salaries, this is a huge tax increase. It will kind of even out over time, but it wacks new companies very hard.
> For software companies, where costs are basically eng salaries,
... it smells like this might be fallout from mis-classifying workers and/or fudging categorization of labor for some benefit. Am I on to something?
The result has been things that cost money are "balanced" by raising taxes somewhere--but politicians don't want to raise taxes. Thus we get all sorts of garbage that fiddles with the details without "raising" taxes, but "raises" revenue--often by pulling it forward rather than actually changing the total amount.
We have also seen a lot of things that employers used to simply pay changed to income for the employee but deductible--but that causes the FICA taxes to be paid in all cases and since an awful lot of employees aren't in a position to itemize those deductions are lost. Something that was tax free now becomes income, but they didn't "raise" taxes.
I'd like to take the idiotic idea and stand it on it's head: I would not permit *any* measure to fund itself. A measure would either be a tax bill or a spending bill, it would be prohibited for a bill to do both. That would remove much of the drive to create insanities like this and Congress could work on cleaning up all the garbage. To accomplish this, though, we will have to evict all those idiots who "promised" never to raise taxes (but are perfectly willing to vote for stealth increases that cause a lot more pain per $ raised than doing it honestly would.)
Which means that at the very least, companies should be able to classify at least some portion of salary costs as "not software development". Maintenance, bug fixing, useless meetings, etc?
It would certainly be consistent with the spirit of R&E to not classify maintenance and bug fixes as R&E, and it would definitely reduce the sting of this change for established companies. Startups would still be pretty screwed.
One could probably apportion some blame to the businesses who assumed a fix from Congress would be forthcoming, but on the whole it seems to me like a spectacularly il-conceived bit of the tax code that never should have been passed in the first place.
Amortizing salaries seems really weird since they are recurring every year. After 5 years you can deduct your full salary expenses for that year. And after you have laid off everybody you can deduct for a few more years. Definitely makes it hard to hire a lot of people quickly if you don't have a ton of profit.