Correct. The most obviously wrong exclusion is stock compensation. Imagine Uber issued and sold $1B of stock each year. Obviously we wouldn’t count that as $1B of profit, because it’s just a change in capital structure. Stock compensation is the same thing. The fact that the stock issuance is bundled together with employee compensation (where it substitutes for cash) instead of going directly to the corporate treasury doesn’t matter.
Well, I think the point is that issuing more stock already shows up in the earnings per share data so there is not a good reason to double count it.
A company that issues stock to employees is not losing money, what it's doing is transferring some wealth from one group of shareholders to a different group of shareholders.
This hurts the bottom line of the first group and helps the bottom line of the second, but it does not hurt the balance sheet of the company.
Likewise if a company votes to give a special dividend to a preferred stock and reduce the dividend to common stock by the same amount, then this hurts one group of shareholders over another, but it does not cause the company to incur a loss.
And as the company's balance sheet is unchanged. its income statement should not be changed either.
The flipside is to believe that the balance sheet should be changed because the liability increases by the amount of new stock. This is false because the liability has to be valued at market value and shareholder issuance only dilutes the value of an individual stock, it does not increase the shareholder equity.
This treatment where GAAP rules misvalue the change to liability from share issuance rears its head when valuing share buybacks. Here too, they are valued at the purchase price, which creates all sorts of misvaluation problems for firms and creates the opportunity for mismanaging of corporate funds by purchasing shares high and selling them low, which really should result in a loss for the company, but according to the GAAP rules there is no loss because the equity is valued by the purchase price/price at grant and does not take into account the market price.
So from this one bad rule comes other misvaluations that distort the balance sheet of a company and impede financial analysis. And all of this stems from a desire to discourage corporate behavior in share issuance rather than from a desire to accurately value the company.
Suppose Uber issued $1B of stock, sold the stock on the open market, then gave employees a $1B bonus. Did Uber incur a $1B expense by paying the bonus, and if so, did they also earn an offsetting $1B in revenue by selling this stock, so this set of transactions should net to zero on the earnings statement? Isn’t this the same thing as stock compensation with an extra step? What am I missing here?
Selling stock doesn't impact the income statement; it is NOT revenue. It is capital raised. Your series of transactions would net to debit expense $1B, credit equity capital $1B. It would definitely fuck up your earnings, as it should.
Stock compensation is compensation and should be expensed. It's not just GAAP, it is plain common sense.
The dilutive effect on EPS isn't any kind of double-counting. It's just a proper accounting for the way the shareholders are constantly bleeding value to sustain normal operations.
As you know, the sale of stock is a capital account operation, so just giving the money to an employee appears to be a current expense with no offsetting current revenue. It would be like selling off a share of your future profits to make current payroll (which is what it is).
But what you are missing is that many modern companies, and especially software companies (2/3 of labor expenses are R&D), have capital that is basically IP created by their employees, and so just as it's a good idea to sell equity to buy some metal stamping machine, and this does not dilute the equity of the firm or cheat any shareholders, so it's also a good idea to sell equity to hire the person to write your application, or create a vaccine, or design a new engine, and this also does not dilute equity or hurt the company. In fact many times when key employees leave the share price of a company falls and the shareholders are more than happy to vote for boards that give stock grants to key employees that cause the company to earn more long term profits.
And ultimately this is something that GAAP just can't handle due to its antiquated distinction between labor and capital investment -- under GAAP you can't expense a labor R&D expense over a period of time. You have to take the charge all in the same year. So this gives you misleading accounting results, and this is also why savvy investors look at the non-GAAP results and then they have to use their judgement to determine whether the IP that the company's employees are creating will generate the expected long term returns. Moreover if you are such a company, it makes excellent sense to finance R&D by selling stock, and for most companies R&D is almost entirely labor compensation.
So ultimately GAAP is just out of touch with what capital investment means in the modern world - capital investment happens by paying people a lot of money to create IP for you, and not by purchasing tractors. Similarly stock analysts who view labor as purely a current expense and not an investment are also out of touch with many industries. That wont work in Biotech or Software or SpaceX or really any firm that does a lot of R&D. So let the stock market decide whether those equity sales are funding compensation that will generate the return or whether they will merely dilute -- the investors are the ones who need to price this, it can't be mechanically determined via GAAP rules designed to track the balance sheet of a 19th century coal plant.
> But what you are missing is that many modern companies, and especially software companies (2/3 of labor expenses are R&D), have capital that is basically IP created by their employees, and so just as it's a good idea to sell equity to buy some metal stamping machine, and this does not dilute the equity of the firm or cheat any shareholders, so it's also a good idea to sell equity to hire the person to write your application, or create a vaccine, or design a new engine, and this also does not dilute equity or hurt the company.
This is a useful way to think about R&D expenses, but it’s not a justification for different treatment of cash and stock comp for employees. It’s more like an argument that R&D investment should be capitalized - and it sometimes is. But whether you capitalize R&D shouldn’t depend on what portion of the expense is stock vs cash.
> but it’s not a justification for different treatment of cash and stock comp for employees.
It is, because the firm will give stock comp -- which is a share of future profits, in proportion to the capital provided by employees -- if it doesn't the shareholders can police that as it will be seen via lower long term P/E ratios. Therefore you really only need the P/E ratios to guide you, you don't need anything else. There is simply no need for GAAP to insist that employee comp is always a current expense and is never a capital investment. That's the issue.
> It’s more like an argument that R&D investment should be capitalized - and it sometimes is
Tech firms can't expense worker comp as capital investment under GAAP. But worker comp accounts for most R&D. Thus tech can't really expense R&D under GAAP properly. That's the problem. That's why these software firms have to show huge GAAP losses when really they are making investments and show non-GAAP profits. It would be the same thing as if capital equipment purchase cost had to be on the income statement and could not be depreciated -- you'd also see huge losses in the first few years followed by absurd earnings. This is exactly the situation tech firms are in, and it explains why both investors and firms focus on non-GAAP earnings in this sector.
What we need is the concept of capital expenditures funded by employee compensation, because the reality of how modern firms work. That's all that's going on here when software firms give stock options to key employees that develop the applications they sell. Instead, GAAP is biased towards the purchase of things like computers and equipment, which is actually a much less useful measure of the true capital that a tech firm has.
At Uber, a large portion of compensation for engineers is in stock. At Netflix, it’s almost entirely cash. Why should non-GAAP earnings privilege Uber’s approach over Netflix’s approach? Wouldn’t it be better to make your non-GAAP earnings capitalize R&D expenses? Even at a company like Uber, plenty of stock goes to non-R&D employees, and plenty of cash does go to R&D, so it’s not even measuring what you want it to measure. Also, capitalizing R&D would mean that companies could also write down failed R&D investment, which is obviously necessary for accurate reporting if you never charged that investment as an expense.
> A company that issues stock to employees is not losing money, what it's doing is transferring some wealth from one group of shareholders to a different group of shareholders.
They are losing money via dilution for the shareholders at the time that they issued the stock.
> They are losing money via dilution for the shareholders at the time that they issued the stock.
I'm not sure what you mean by "losing money via dilution". Companies don't lose money via the market repricing their shares. Shareholders may lose money. But in aggregate, there are also more shareholders, so one group of shareholders loses some money and another group of shareholders gains some money. In aggregate, shareholders don't lose money.
Happily, the FASB isn't fooled by this particular sleight-of-hand. Not anymore, at any rate; they quit letting themselves get bamboozled by Silicon Valley flim-flam accounting for options more than 15 year ago.
If the shareholders have constant ongoing dilution due to continual share issuance, they have run just to stay in place. This is pretty straightforward, and what the GAAP are communicating to FS readers. You have a comp expense, of course (you gave something valuable to an employee instead of selling it for cash), and you also have the dilution, without even getting the cash in return. It's the only appropriate way to account for it.
It is not flim-flam. The shareholders elect a board, and the board has to approve stock based compensation. The shareholders price shares based on this expected dilution so no one is being fooled and there is no sleight of hand. All of this is very public.
The flim-flam comes when management argues that they don't need to book a compensation expense. Happily, though, this flaw in the GAAP has long since been cured.
The biggest chunk was excluding their net loss in investments in other companies.
However, "adjusted earnings" is a PR number it doesn't have any particular financial meaning. If Uber wanted to make the case that they are moving towards profitability, they could talk about operating income or cash flow from operations or some other metric that has a GAAP meaning.
Those are the numbers to watch if you want to know if the company can ever be profitable.
Chinese Government announced it was investigating Didi for national security, which is usually code for corruption/ scaring management into greater compliance.
Didi was a massively valuable Chinese company Uber owned a significant amount of. When the stock value crashed (-23% this quaarter), it destroyed a significant amount of book-value for uber who owned it.
That’s true for most adjustments. Buffett also suggests investors ignore income statement effects of marking securities to market.
“I must first tell you about a new accounting rule – a generally accepted accounting principle (GAAP) – that in future quarterly and annual reports will severely distort Berkshire’s net income figures and very often mislead commentators and investors.
The new rule says that the net change in unrealized investment gains and losses in stocks we hold must be included in all net income figures we report to you. That requirement will produce some truly wild and capricious swings in our GAAP bottom-line.”
https://www.berkshirehathaway.com/2017ar/2017ar.pdf
EBITDA is bullshit if depreciation is an expected, significant part of your business. For many vc-backed companies, it isn't, but for any company with major physical capex, it very much is a real expense.
There's also the whole inconsistency with capitalizing software expenses... if you want to be aggressive about accounting, you certainly can.
But of course the biggest part of this story is stock-based comp, which is very much a real expense.
(1) uber owns nothing (mostly), and
(2) depreciation doesn't really affect real-world profit/loss as a "regular person" expects. Its mostly an accounting tool (that often lets companies hide money). If you eg. build a building, you can depreciate ~10% a year in many cases, but that doesn't really mean you're losing/gaining money.
Eg. a residential building can depreciate at 5% a year. If you owned and rented a house in SF, its not decreasing in value by 5% a year, probably even considering cost of repairs/upgrades grows as a building ages (the purpose of this metric). So including a (-5% of building) depreciation amount on the earnings of said rental property makes little sense when evaluating if its "profitable".
“Uber’s preferred “adjusted ebitda” metric of the company’s underlying health excludes multiple costs including interest, taxes, depreciation and amortisation, as well as stock-based compensation.”
> excludes multiple costs including interest, taxes, depreciation and amortisation
That is exactly what "EBIDTA" means. And EBIDTA is pretty standard. The "IDTA" is not nearly as useful to gauge the health of a company as EBIDTA.
Eg. companies that don't run a profit don't pay taxes, so if you make $0 after taxes, you basically have your whole tax amount to spend next year and it makes no difference.
> [EBITDA] is derived by subtracting from revenues all costs of the operating business (e.g. wages, costs of raw materials, services ...) but not depreciation, amortisation, interest, lease expenses, and taxes.
(1) i've recently heard a lot of people complain about price increases, so it's worrying to think a service many people using would need to be a lot more expensive
(2) it validates a lot of the business models for start ups to get big to find profits, and validates that losing money - even lots - can lead to profits and hopefully sustainability.
doesn't really seem like a healthy/fair market if the winner is picked by incumbent capitalists
also you forgot a key part of the equation, which was to ignore pesky regulatory requirements
I would hope startups aren't encouraged to follow this model but I hear "Uber for X" way too often so I know it already doesn't matter if Uber ever makes (non-VC) money, as the pyramid scheme is already seen as a wild success.
Are they profitable? No, but maybe closer than some are thinking, and further than others lol.
If you ignore their 1.8B loss on Other Income which I think is mostly Didi, you've got a few big items left that seem likely to be recurring and "important":
* + 100M gain on taxes (yes, gain as in benefit)
* - 100M loss on interest
* - 200M loss on depreciation
* - 300M loss on stock comp
So if I were to adjust their adjusted number, I would say they're still about 500M in the red for the quarter. In the most generous of takes I'd say they can get 40% gross margin on new revenue. So to make up $500M they'll need to do $1.25B more revenue which is about 26%. So... yeah I guess they could get there.
38 comments
[ 35.9 ms ] story [ 186 ms ] threadA company that issues stock to employees is not losing money, what it's doing is transferring some wealth from one group of shareholders to a different group of shareholders.
This hurts the bottom line of the first group and helps the bottom line of the second, but it does not hurt the balance sheet of the company.
Likewise if a company votes to give a special dividend to a preferred stock and reduce the dividend to common stock by the same amount, then this hurts one group of shareholders over another, but it does not cause the company to incur a loss.
And as the company's balance sheet is unchanged. its income statement should not be changed either.
The flipside is to believe that the balance sheet should be changed because the liability increases by the amount of new stock. This is false because the liability has to be valued at market value and shareholder issuance only dilutes the value of an individual stock, it does not increase the shareholder equity.
This treatment where GAAP rules misvalue the change to liability from share issuance rears its head when valuing share buybacks. Here too, they are valued at the purchase price, which creates all sorts of misvaluation problems for firms and creates the opportunity for mismanaging of corporate funds by purchasing shares high and selling them low, which really should result in a loss for the company, but according to the GAAP rules there is no loss because the equity is valued by the purchase price/price at grant and does not take into account the market price.
So from this one bad rule comes other misvaluations that distort the balance sheet of a company and impede financial analysis. And all of this stems from a desire to discourage corporate behavior in share issuance rather than from a desire to accurately value the company.
So I'll go by the non-GAAP figures.
Stock compensation is compensation and should be expensed. It's not just GAAP, it is plain common sense.
The dilutive effect on EPS isn't any kind of double-counting. It's just a proper accounting for the way the shareholders are constantly bleeding value to sustain normal operations.
But what you are missing is that many modern companies, and especially software companies (2/3 of labor expenses are R&D), have capital that is basically IP created by their employees, and so just as it's a good idea to sell equity to buy some metal stamping machine, and this does not dilute the equity of the firm or cheat any shareholders, so it's also a good idea to sell equity to hire the person to write your application, or create a vaccine, or design a new engine, and this also does not dilute equity or hurt the company. In fact many times when key employees leave the share price of a company falls and the shareholders are more than happy to vote for boards that give stock grants to key employees that cause the company to earn more long term profits.
And ultimately this is something that GAAP just can't handle due to its antiquated distinction between labor and capital investment -- under GAAP you can't expense a labor R&D expense over a period of time. You have to take the charge all in the same year. So this gives you misleading accounting results, and this is also why savvy investors look at the non-GAAP results and then they have to use their judgement to determine whether the IP that the company's employees are creating will generate the expected long term returns. Moreover if you are such a company, it makes excellent sense to finance R&D by selling stock, and for most companies R&D is almost entirely labor compensation.
So ultimately GAAP is just out of touch with what capital investment means in the modern world - capital investment happens by paying people a lot of money to create IP for you, and not by purchasing tractors. Similarly stock analysts who view labor as purely a current expense and not an investment are also out of touch with many industries. That wont work in Biotech or Software or SpaceX or really any firm that does a lot of R&D. So let the stock market decide whether those equity sales are funding compensation that will generate the return or whether they will merely dilute -- the investors are the ones who need to price this, it can't be mechanically determined via GAAP rules designed to track the balance sheet of a 19th century coal plant.
This is a useful way to think about R&D expenses, but it’s not a justification for different treatment of cash and stock comp for employees. It’s more like an argument that R&D investment should be capitalized - and it sometimes is. But whether you capitalize R&D shouldn’t depend on what portion of the expense is stock vs cash.
It is, because the firm will give stock comp -- which is a share of future profits, in proportion to the capital provided by employees -- if it doesn't the shareholders can police that as it will be seen via lower long term P/E ratios. Therefore you really only need the P/E ratios to guide you, you don't need anything else. There is simply no need for GAAP to insist that employee comp is always a current expense and is never a capital investment. That's the issue.
> It’s more like an argument that R&D investment should be capitalized - and it sometimes is
Tech firms can't expense worker comp as capital investment under GAAP. But worker comp accounts for most R&D. Thus tech can't really expense R&D under GAAP properly. That's the problem. That's why these software firms have to show huge GAAP losses when really they are making investments and show non-GAAP profits. It would be the same thing as if capital equipment purchase cost had to be on the income statement and could not be depreciated -- you'd also see huge losses in the first few years followed by absurd earnings. This is exactly the situation tech firms are in, and it explains why both investors and firms focus on non-GAAP earnings in this sector.
What we need is the concept of capital expenditures funded by employee compensation, because the reality of how modern firms work. That's all that's going on here when software firms give stock options to key employees that develop the applications they sell. Instead, GAAP is biased towards the purchase of things like computers and equipment, which is actually a much less useful measure of the true capital that a tech firm has.
They are losing money via dilution for the shareholders at the time that they issued the stock.
I'm not sure what you mean by "losing money via dilution". Companies don't lose money via the market repricing their shares. Shareholders may lose money. But in aggregate, there are also more shareholders, so one group of shareholders loses some money and another group of shareholders gains some money. In aggregate, shareholders don't lose money.
Happily, the FASB isn't fooled by this particular sleight-of-hand. Not anymore, at any rate; they quit letting themselves get bamboozled by Silicon Valley flim-flam accounting for options more than 15 year ago.
If the shareholders have constant ongoing dilution due to continual share issuance, they have run just to stay in place. This is pretty straightforward, and what the GAAP are communicating to FS readers. You have a comp expense, of course (you gave something valuable to an employee instead of selling it for cash), and you also have the dilution, without even getting the cash in return. It's the only appropriate way to account for it.
It is a very expensive form of compensation.
As with any other way of spending money, that is often in their shareholders' best interest, but arguing that it is not an expense is nonsense.
However, they posted an actual Q3 loss of $2.4B.
On an adjusted basis I look just like George Clooney.
However, "adjusted earnings" is a PR number it doesn't have any particular financial meaning. If Uber wanted to make the case that they are moving towards profitability, they could talk about operating income or cash flow from operations or some other metric that has a GAAP meaning.
Those are the numbers to watch if you want to know if the company can ever be profitable.
They bought some stake in company Didi.
In this past quarter, they acknowledged this stake was worth 3.2b less than it was recorded as in their books.
Therefore, their books now "lost" 3.2B in value.
But they did not "overspend" by 3.2B in the sense most people expect when hearing "they just lost 3.2B this quarter"
Lol, how does that happen, honestly.
Chinese Government announced it was investigating Didi for national security, which is usually code for corruption/ scaring management into greater compliance.
Didi was a massively valuable Chinese company Uber owned a significant amount of. When the stock value crashed (-23% this quaarter), it destroyed a significant amount of book-value for uber who owned it.
“I must first tell you about a new accounting rule – a generally accepted accounting principle (GAAP) – that in future quarterly and annual reports will severely distort Berkshire’s net income figures and very often mislead commentators and investors. The new rule says that the net change in unrealized investment gains and losses in stocks we hold must be included in all net income figures we report to you. That requirement will produce some truly wild and capricious swings in our GAAP bottom-line.” https://www.berkshirehathaway.com/2017ar/2017ar.pdf
No. Adjusted EBITDA. Meaning not EBITDA. EBITDA is not "bullshit earnings".
There's also the whole inconsistency with capitalizing software expenses... if you want to be aggressive about accounting, you certainly can.
But of course the biggest part of this story is stock-based comp, which is very much a real expense.
(1) uber owns nothing (mostly), and (2) depreciation doesn't really affect real-world profit/loss as a "regular person" expects. Its mostly an accounting tool (that often lets companies hide money). If you eg. build a building, you can depreciate ~10% a year in many cases, but that doesn't really mean you're losing/gaining money.
Eg. a residential building can depreciate at 5% a year. If you owned and rented a house in SF, its not decreasing in value by 5% a year, probably even considering cost of repairs/upgrades grows as a building ages (the purpose of this metric). So including a (-5% of building) depreciation amount on the earnings of said rental property makes little sense when evaluating if its "profitable".
In some countries development costs can be captalised, and the depreciated over future years (when that software is earning revenue)
Sort of ignores the fact that software is always a depreciating asset from the time you first start writing it
Thats Operational Expenditure (OpEx) not Capital Expenditure (CapEx) usually. Unless they own servers, or bought licenses to software, etc.
See https://www.ft.com/content/4af56131-018a-412e-93d9-dcbb37232...
That is exactly what "EBIDTA" means. And EBIDTA is pretty standard. The "IDTA" is not nearly as useful to gauge the health of a company as EBIDTA.
Eg. companies that don't run a profit don't pay taxes, so if you make $0 after taxes, you basically have your whole tax amount to spend next year and it makes no difference.
> [EBITDA] is derived by subtracting from revenues all costs of the operating business (e.g. wages, costs of raw materials, services ...) but not depreciation, amortisation, interest, lease expenses, and taxes.
https://en.wikipedia.org/wiki/Earnings_before_interest,_taxe...
Its great both because
(1) i've recently heard a lot of people complain about price increases, so it's worrying to think a service many people using would need to be a lot more expensive
(2) it validates a lot of the business models for start ups to get big to find profits, and validates that losing money - even lots - can lead to profits and hopefully sustainability.
also you forgot a key part of the equation, which was to ignore pesky regulatory requirements
I would hope startups aren't encouraged to follow this model but I hear "Uber for X" way too often so I know it already doesn't matter if Uber ever makes (non-VC) money, as the pyramid scheme is already seen as a wild success.
Here are the actual income statements: https://investor.uber.com/news-events/news/press-release-det...
Are they profitable? No, but maybe closer than some are thinking, and further than others lol.
If you ignore their 1.8B loss on Other Income which I think is mostly Didi, you've got a few big items left that seem likely to be recurring and "important":
* + 100M gain on taxes (yes, gain as in benefit)
* - 100M loss on interest
* - 200M loss on depreciation
* - 300M loss on stock comp
So if I were to adjust their adjusted number, I would say they're still about 500M in the red for the quarter. In the most generous of takes I'd say they can get 40% gross margin on new revenue. So to make up $500M they'll need to do $1.25B more revenue which is about 26%. So... yeah I guess they could get there.