EV/FCF ratio...a superior valuation metric?

EV/FCF vs. P/E

Over the past few weeks, I have been thinking a decent amount about Enterprise Value (EV) being far superior for share price valuation considerations than market cap (the P in P/E or P/S ratios). While P/Es get thrown about all over the place and could probably be explained by relatively beginner investors, the use of market cap (the price in P/E) doesn’t account for the ownership dilution due to bondholders having a right to part of a company’s future cash flows (i.e. a net debt position). The converse side of that is that for a net cash position, the debt could be instantly paid off in theory, leaving the entire net cash position as owned solely by (or theoretically distributable to) the shareholders. The simple shift of using EV rather than market cap accounts for this/these factor(s).*

Along with this line of thinking** and with Tinker’s observation of Arista’s present EV/FCF ratio (linked below*** & ****), I have started thinking that EV/cash flow might ultimately be a superior way to determine if companies are overvalued/undervalued compared to P/E (lower ratio generally indicating more undervalued…and lower ratios eventually being corrected by share price rises).

Similarly to PEG “ratios” (or 1YPEG), an estimated future growth rate of the cash flow could be used in an EV/CF/G ratio. The investor’s judgment of the company’s competitive advantage period (CAP****) could then factor into their estimate of the future direction of cash flow growth rate.

I will refrain from going in-depth about cash flow being superior to GAAP earnings for this particular valuation metric, but that could be worth some decent discussion as well. I have yet to test my hypothesis, and also have not yet convinced myself whether simple operating cash flow (OCF) is sufficient for this metric or whether free cash flow (FCF) would be substantially better to the point of warranting calculating FCF rather than simply reading OCF from the cash flow statement.

Ultimately though, use of this method would provide a fair correlation to using a discounted cash flow valuation (often considered the true essence of valuing a business)…while also properly accounting for the percentage ownership of shareholders/debt-holders and giving proper credit for cash that has already been generated and is still maintained by the company.

Last night, Tinker pointed out that with Arista’s $1.5 billion net cash position, ANET is at about a 22.5 on an EV/FCF basis (forward FCF, based simply on Q4 17 x 4). That sounds quite a bit less overvalued than the P/E ratio for ANET…even with “only” about a 35% revenue growth rate. Tinker actually just put up a new post on the NPI board further expounding on his thoughts about ANET’s present valuation, which pairs nicely with my thoughts here (which his posts had influenced)****.

-volfan84
Long ANET

  • Also as a small note, if a company has preferred shares, holders of those get paid/have a right to the cash before the holders of common shares as well, but that isn’t something that gets encountered much with stocks that are covered on this board (from what I can tell).

** This thinking was sparked in part by hearing in late January about Amazon passing Apple on an Enterprise Value basis and also by noting Bert’s frequent usage of EV/Sales ratios in his write-ups.

*** http://discussion.fool.com/arista-analysis-32987839.aspx?sort=wh…

**** http://discussion.fool.com/final-thoughts-on-arista-32988609.asp…

***** Bear’s marginal thinking posts have also factored into my brain traveling down these lines of thinking

20 Likes

Not to be skeptical or anything … :slight_smile: … but the first thing I would do with a proposed metric like this is to apply it to a bunch of companies and see whether it fits with my opinion based on other known information. This requires a spectrum because, if one applies it only to companies that one already likes, there is a strong danger of confirmation bias, i.e., a good company is going to look good over a wide range of metrics. At the very least, a bad company should look bad and a marginal one might look better or worse … and one would learn something. After all, if one invents metrics which do no more than confirm what we thought already, then it has added nothing to our knowledge except the illusion of having a number. If, on the other hand, it helps to distinguish between good and bad options among companies that we were unsure about, it has added value.

8 Likes

a superior valuation metric?

Every day I’m more convinced that traditional valuation metrics used in “Security Analysis” are not applicable to Saul type stocks. These metrics were devised for the industrial revolution type companies like assembly plants and steel mills and for service companies like department stores. All these businesses have one feature in common, they are “decreasing returns” type businesses. To sell more in department stores you have to build mores stores. To make more steel you need more smelters. To make more cars you need more assembly lines. By contrast the kinds of businesses Saul is investing in currently are “increasing returns.” To sell more software all you need is to have more customers download the code, you don’t need more lines of code or more coders. You might need more help desk personnel. Another important difference is that with increasing returns type businesses the winner tends to take most of the market share. Applying steel mill logic to NVDA simply makes no sense even if you multiply the P/E ratio by five.

Tinker pretty much nails it in the post you linked:

One thing we can quantify is if the stock really is as expensive as it is perceived to be by those using conventional metrics. My answer is a clear, no, not at all. Sure, price to sale, price to earnings. But none of that matters to your bank account. What matters to your bank account is what goes into it, not how expensive the source of that money is perceived to be.

What makes the price of the stock go up? Growth! In the same post Tinker warns about bubbles. When the market gets too excited the latecomers pay too much. Growth in nature has been studied quite a bit and there is a definite pattern to it best represented by the “S” curve. In general terms divide the lifetime of an industry or of a technology (not of companies or stocks) into three equal length periods. The first third is slow growth as the technology tries to establish a beachhead in the market. Most fail. When around 15% market penetration is achieved and people start seeing the benefits, growth accelerates creating the “curve in the hockey stick.” Fast growth continues until around 85% market penetration. As you run out of new customers to sell to the top of the “S” curse sets in. At that point the industry can be a very profitable cash cow but with little growth and standard valuation metrics again apply. Qualcomm, Cisco and Microsoft are good examples.

Non of this is smooth sailing. During the fast growth middle period it is quite common to see a stock lose 50% of its value only to get back on track. You have to check for the disruptors Tinker talks about about.

Once you have established that the technology a company is selling is an “increasing returns” winner the most important metric is market penetration. The biggest danger from the outside is disruption and the biggest danger from the inside is bad management, spending too much to buy growth. You want a product that people want to buy, not one you have to push to sell. :wink:

Two comments on GAAP earnings and how they distort reality, stock based compensation and expensing R&D.

The powers that be (Buffett, FASB, et al) take a dim view of stock based compensation but it is very beneficial for shareholders provided it is not abused. Let’s start from the proposition that there is no free lunch. If you run a steel mill your biggest fixed cost the plant and equipment. If you make GPU chips your biggest fixed cost if the army of engineers. Let’s suppose you can either pay them a fixed salary of $3 million or $1 million plus stock options that generate an extra $4 million for each engineer. If you have 5000 engineers the fixed salary method requires a $15 billion payroll. With stock options your payroll is reduced to $5 billion. This gives the company a $10 billion cushion. It might not matter to a mature cash cow but it most certainly matters to a startup that has not yet reached the magical 15% market penetration. Of course there is no free lunch, shareholders pay the engineers with dilution but without the engineers there would be no company. Anyone here refuses to buy ANET or NVDA because they have stock options? Stock options are a very smart way to run a risky increasing returns people business. Don’t let the conventional wisdom blind you to the benefits of well designed stock option programs.

From a traditional P/E valuation point of view, how much would the P/E ration drop if you had to add an extra $10 billion to expenses or to reduce after tax net profit by $7 or $8 billion?

Expensing R&D has a similar but inverse effect on traditional metrics. By expensing R&D instead of capitalizing it, you increase costs and reduce profits but at the same time the accounting hides the value of the software being developed, software that will generate increasing amounts of revenue in the future. Software is the present day equivalent of 20th century steel smelters but we account for it differently which I consider is a big mistake.

From this point of view, how much relevance does Shiller CAPE have today?

Denny Schlesinger

55 Likes

Expensing R&D is a great point Denny. An easy example is the $3 to $5 billion, probably the latter, that Nvidia spent to create the Volta with tensor cores. That R&D is expensed as one time w current accounting practices, but in reality, as you say, it is a substitute for the smelting machine. The smelting machine is depreciated over its useful life. The Volta’s useful life is what, 5 or 10 years? To be an accurate assessment it should be depreciated over that period of time, at least. As the research on the Volta will be leveraged to create the next chip and so on.

This is one reason why tech companies have higher valuations than traditional companies, and those valuations have remained through decades despite fundamentalists decrying this. 1 year expensing R&D is one reason, in low capital, increasing return businesses, that earnings for tech companies are artificially low from an accounting perspective.

Arista has IP in technology that will probably last at least a decade or longer.

The problem is, is that it is quite subjective as to the useful life of the R&D, and the research can be disrupted at any point in time.

As such, I think we can comfortably live with higher valuations in low capital, increasing return companies, and more so with companies with large CAPs, as the R&D investments are more likely to have longer and undisputed useful lives to depreciate against.

But a lot of research and development does not pan out, and therefore has no useful life, and a lot of research and development are things like what SHOP does, adding features here and features there, where it is impossible to understand, except perhaps with some sort of regression analysis, if one wanted to try, to figure out what these features are actually marginally worth per year.

And you know, this is probably the same discussions that go on with cost accounting conventions. Gee, no so boring after all! Oh, yes it is. But your in Vegas! Yeah, I did a marketing convention in Vegas - no comment.

Tinker

15 Likes

The problem is, is that it is quite subjective as to the useful life of the R&D, and the research can be disrupted at any point in time.

But a lot of research and development does not pan out, and therefore has no useful life, and a lot of research and development are things like what SHOP does, adding features here and features there, where it is impossible to understand, except perhaps with some sort of regression analysis, if one wanted to try, to figure out what these features are actually marginally worth per year.

For an accountant none of the above should be a problem that can’t be dealt with. Management should have enough discretion to set a useful-life number to any R&D project. If the R&D does not pan out you expense whatever value the asset still has just as is done with goodwill impairment.

The problem is the reverse! Expensing R&D increases cost and lowers taxes, an incentive to expense instead of to capitalize. The opposing incentive is to show higher profit but if your company is valued more by cash flow than by earnings this incentive is negated. Bezos has been a master at growing Amazon while showing no profits. Where the heck did all the money come from to grow Amazon if not from profits? It’s all an accounting mirage.

My point, really, is that metrics like Shiller CAPE do NOT apply to high tech.

Denny Schlesinger

6 Likes

Every day I’m more convinced that traditional valuation metrics used in “Security Analysis” are not applicable to Saul type stocks. These metrics were devised for the industrial revolution type companies like assembly plants and steel mills and for service companies like department stores. All these businesses have one feature in common, they are “decreasing returns” type businesses. To sell more in department stores you have to build mores stores. To make more steel you need more smelters. To make more cars you need more assembly lines. By contrast the kinds of businesses Saul is investing in currently are “increasing returns.” To sell more software all you need is to have more customers download the code, you don’t need more lines of code or more coders…What makes the price of the stock go up? Growth!..

Denny,
What a great post. Thanks so much.
Saul

10 Likes

Expensing R&D is a great point Denny. An easy example is the $3 to $5 billion, probably the latter, that Nvidia spent to create the Volta with tensor cores. That R&D is expensed as one time w current accounting practices, but in reality, as you say, it is a substitute for the smelting machine. The smelting machine is depreciated over its useful life. The Volta’s useful life is what, 5 or 10 years? To be an accurate assessment it should be depreciated over that period of time, at least. As the research on the Volta will be leveraged to create the next chip and so on.

This is one reason why tech companies have higher valuations than traditional companies, and those valuations have remained through decades despite fundamentalists decrying this. 1 year expensing R&D is one reason, in low capital, increasing return businesses, that earnings for tech companies are artificially low from an accounting perspective. Arista has IP in technology that will probably last at least a decade or longer…

I love that too, Tinker. Yes, the R&D for new software or a new product in these companies is the equivalent of the expense of building new manufacturing machines or a new factory for a manufacturing company.
Best,
Saul

10 Likes

Every day I’m more convinced that traditional valuation metrics used in “Security Analysis” are not applicable to Saul type stocks. These metrics were devised for the industrial revolution type companies like assembly plants and steel mills and for service companies like department stores. All these businesses have one feature in common, they are “decreasing returns” type businesses. To sell more in department stores you have to build mores stores. To make more steel you need more smelters. To make more cars you need more assembly lines. By contrast the kinds of businesses Saul is investing in currently are “increasing returns.” To sell more software all you need is to have more customers download the code, you don’t need more lines of code or more coders.

Hi again Denny,

I was thinking that my thoughts from my recent post on TLND (#37427) fits in with what you wrote. I’m reposting it here a little expanded:

The difference is that if you are selling refrigerators, next year you have to sell them all over again to a different set of customers. On the other hand, a SaaS company sells software to a new customer and then has that same subscription revenue, and usually more, every quarter…forever. They are well aware of this and thus feel that it makes sense for them to grab every customer they can, while they can, even if it means plowing back all their gross margin profits into sales and marketing. I agree. They pay those marketing commissions this quarter, but they are only balanced by this single quarter’s subscription payment in the quarterly report. They may thus show no profit, or even a loss on the initial quarter’s sales, depending on commissions. (Thus a quarter with huge sales can even, paradoxically, show a lower operating margin percent). But, they’ll recognize another subscription payment next quarter, and the one after, and the one after, and the one after…without the commissions. And a company with a 125% retention rate actually has 25% more revenue from that sale the second year, and even 25% more the next year.

Saul

18 Likes

I was thinking that my thoughts from my recent post on TLND (#37427) fits in with what you wrote. I’m reposting it here a little expanded:

Saul, similar but not quite the same. It is the business model IBM used until the consent decree of 1950 something. In the early days IBM didn’t sell their machines, they rented them on a monthy basis. The consent decree forced IBM to either rent or sell depending on the customers’ wishes.

There was a difference from what you describe when I was selling IBM equipment. We received commissions on a monthly basis but repeat business didn’t count toward the sales quota, only new or additional billing counted. After all, we had the job of keeping our IBM customers happy and upgrading them when possible. Are you sure Talend only pays commissions on new sales? Does not make a lot of sense to me from my past experience. If they can get away with it the business model is even more aggressive than IBM’s was.

Denny Schlesinger

Dated 05.02.97
JUDGE TO IBM: IT’S NOT 1956 ANYMORE

BIG BLUE JUST ain’t what it used to be. International Business Machines announced Friday that a federal judge has approved an agreement with the Justice Department ending a 1956 decree that limited the company’s clout in the then-burgeoning computer industry. “The situation has changed drastically in the 40 years since the time the consent decree was entered,” Judge Thomas Griesa wrote. “It has been established beyond any real question that, whereas IBM formerly had a great degree of market power in an antitrust sense, that market power has been substantially diminished, and is continuing to diminish, to the point of disappearance in the sense of a threat of antitrust violation.”

https://www.wired.com/1997/05/judge-to-ibm-its-not-1956-anym…

7 Likes

how much relevance does Shiller CAPE have today? most companies are more like a grocery store than a software company. CAPE applies to broad indices, and just implies probable future return over a multi year span, it can not be used for timing.

Even then I feel it may under-rate the stock market potential because the world is a far safer place than it has been almost any time in the past 100 years, risk is more likely to be rewarded, and technology is compounding at rapid rate. The early 1900’s prior to WW I ,was the last time we had anything similar in the world. But because of the compounding the rate of change has accelerated since then.

Perhaps add to that the ability of governments to pump money into the economy and create huge debt. My guess is that these government actions will end badly but it could be decades. The strong deflating effect of technology is allowing them to get away with this without producing ruinous inflation levels. So far. My guess is that these government actions will end badly but it could be decades away.

Another point about bull and bear markets, They only happen every 4 t 8 years or so, the data base here is small, and taking place over a time of rapid change. So don’t draw too many conclusions. Though I suspect greed fear and panic work about the same now as they did in Imperial Roman grain markets.

Every day I’m more convinced that traditional valuation metrics used in “Security Analysis” are not applicable to Saul type stocks me too. Which gives us an edge.

One other comment “downloading the software” is very quickly done therefore the TALC can be speeded up

3 Likes

Are you sure Talend only pays commissions on new sales? Does not make a lot of sense to me from my past experience. If they can get away with it the business model is even more aggressive than IBM’s was.

Not a clue, Denny. I was just giving a generic example.

I have to wade in with a dissenting opinion about R&D costs.

First, technology advances have a very short life span. Even companies like Nvidia, and ANET that appear ahead of the game, have to keep their pedal to the metal to stay ahead. If they coasted for a year, two max, they will get caught and passed. So, I disagree with concept that some companies have a 5 or 10 asset R&D asset. They may have a product that is ahead, and with prudent management and a lot of R&D will stay ahead, but the shelf life of what they have is much shorter than 5 years. They may stay ahead for 5 years, but I would argue its because of good management, and quality R&D moving forward, not because they are so much superior now they can coast.

Second, how do you determine what R&D has any shelf life, and which just went up in smoke. Allowing management to determine, this amount of this year R&D has future value is just asking for financial statements that can’t be read by anyone outside of the company. Accounting rules are not perfect, and many on this board have pointed out GAAP vs non-GAAP. And generally the outcome of whether R&D has value, takes a while for the sales department to sell their customers on the value of this latest software code. 2018 R&D turns into 2019 sales, and the value of that won’t be determined until 2019. Increasingly, January 2018 R&D turns into April 2018 sales.

Lastly, very few companies really have value in a manufacturing plant. Take for example Ford Motor. Is the value in Ford in the factories, or is it in the technology to build a motor, a transmission, a suspension system, the styling of the car, the seats, the handling, and increasingly, the software that drives all of the above. I would argue is in the technology of not only the cars they sell, the technology of how to bring 5,000 or more different parts together into a car I, you and everyone else wants to buy.

I live in the Seattle area, Boeing country. Boeing has been going great guns for years, and I have avoided buying the stock. Maybe not the smartest move, but until recently they have been capitalizing the costs of building a 787. Under GAAP they build airplanes using a concept called Program Cost (at least I think that is the name). Using this concept, management attempts to figure out the R&D costs of designing the plane, the expected total revenue from the sales of the plane, and the expected total costs to manufacture the plane. Since it takes them a long time to train the staff, fix the manufacturing process to reduce costs etc, GAAP has allowed them to capitalize the cost of building a 787 more than 5 years after they first started building them. Frankly, the CPA in me thinks that is crazy. So, every year Boeing management figures out the new updated total program sales, updated total program cost, and they only expense what management wants to expense. So, until recently they sell a 787 for say $ 20 million, and it takes them $22 million in parts and labor to build it, but they deduct only $16 million on their income statement, and add the difference to their capitalized program costs. I think that is nuts.

And I think you are basically advocating doing the same with R&D.

The way to evaluate R&D value is just exactly like Saul advocates. How are the sales year over year, quarter over quarter. If they are going up, R&D is working, if they are not, I don’t want the R&D going to the balance sheet.

14 Likes

For an accountant none of the above should be a problem that can’t be dealt with. Management should have enough discretion to set a useful-life number to any R&D project.

Actually, Under GAAP, firms are required to expense R&D in the year spent. (http://www.valens-research.com/rd-investment-not-expense-cap… )

For example, if you’re Boeing and you’re spending a lot on R&D for the 787, you end up having a year that looks really bad, followed by years that look too good. Amazon’s hyper growth and its business’s low capital requirements enabled it was able to play this game really well while companies like Tesla are being extra penalized right now.

Software is the present day equivalent of 20th century steel smelters but we account for it differently which I consider is a big mistake…My point, really, is that metrics like Shiller CAPE do NOT apply to high tech.

Baruch Lev, a professor at NYU who has written a book “The End of Accounting.” In it he argues that companies should not only capitalize R&D but separate actual research from development , in essence reinforcing Danny’s point:

The industrial economy, he says, with its emphasis on physical assets of property, plant and equipment, and inventory levels has been supplanted by a new economy in which intangible assets like research and development, information technology, unique business franchises, and powerful brands rule the roost.( https://www.barrons.com/articles/a-new-book-argues-gaap-acco… )

“This is why classical GAAP accounting measures like earnings per share, selling, general, and administrative [SG&A] spending, and return on equity now have lost their relevance and predictive value for investors."

And believe it or not, this change is business type is also related to Stock-Based Compensation. Danny writes:
Stock options are a very smart way to run a risky increasing returns people business… If you run a steel mill your biggest fixed cost the plant and equipment. If you make GPU chips your biggest fixed cost if the army of engineers.If you have 5000 engineers the fixed salary method requires a $15 billion payroll…

That’s all true, but there’s more than just a “pay people later when we’re successful for the work they’re doing today that will actually make us successful” advantage. There’s attraction and retention of engineering talent.

Say you were an engineer 100 years ago and had a better way to smelt steel (btw, I have no idea what that means), if you wanted to start your own business to profit from it, you’d need a lot of capital, which meant loans, which meant showing a business plan in which conservative people like bankers would believe, which essentially meant that unless you had already done that before it was unlikely you’d be able to do it now. But today, thanks to the low overhead of software development, you can actually start writing the software yourself, eventually demoing to VCs who can take riskier bets since the investments are smaller and the payoffs are bigger.

Buffet may not like SBC, but he’s in things like candy, soda, and insurance - hardly the types of business where software is key. His track record in high-tech is pretty bad, too.

So, my tie-in point is that many new businesses not only need better accounting standards to reflect where the real value them lies (people, not equipment), they need to keep those people at their company.

8 Likes

Looks like this metric is used at least in the following site:
https://www.stock-analysis-on.net/NASDAQ/Company/Facebook-In…

I haven’t yet checked the site to see if they have calculated the EV/FCF ratio for other companies.

Certainly, this won’t be a metric to use universally (as P/E also should not be), but it might be useful in recognizing some value that some less sophisticated investors could be spooked out of…and Arista could possibly be a prime example (maybe THE poster child company), but still TBD of course.

-volfan84

Baruch Lev, a professor at NYU who has written a book “The End of Accounting.”

Here is a link to the book.

The End of Accounting and the Path Forward for Investors and Managers (Wiley Finance) https://www.amazon.com/dp/1119191092/ref=cm_sw_r_cp_apa_SAEI…

And I think you are basically advocating doing the same with R&D.

Winlockdon, I’m not sure who you are addressing that to. At this point I’m not advocating anything, what I said was: “Every day I’m more convinced that traditional valuation metrics used in ‘Security Analysis’ are not applicable to Saul type stocks.” While I have studied and practiced accounting I’m not a CPA and I have not thought about reforming GAAP. For ten years I worked as a management consultant and have a good grasp of business models on which I rely heavily in my investment decisions. I currently use the financial statements mostly in the negative, looking for reasons to NOT invest in the company, red flags such as too much debt and excessive SG&A that eats up revenue gains.

My reason for bringing up R&D was to show how high tech is different from old style industrials – not necessarily new style industrials like your Ford Motor example. I’m comparing businesses vs. old style security analysis. While I’m on the subject of R&D I want to point out a curious anomaly.

Suppose Company Developer spends a million dollars on a successful R&D project which Mr. Market values at $5 million market CAP. Company Developer has zero assets, zero book value, having expensed the R&D. Along comes Company Buyer with an offer to buy Company Developer for $5.5 million. If I’m not mistaken, since Company Developer has a book value of zero dollars, Company Buyer has to include $5.5 million as Goodwill assets on the balance sheet. As if by magic, the R&D has been capitalized not to the tune of one but of five point five million dollars! All else being equal, Company Developer and Company Buyer should be identical but are not.

Denny Schlesinger

1 Like

Let’s expand that discussion of R&D expense a bit to include all intellectual property, or if you will, information in general . . .

In my previous life as an Enterprise Architect in a large, multinational aerospace firm I specialized in information management (a thankless task). There are monumental problems in this regard, primary among them is the fact that most firms don’t actually manage their information. They may protect it, pretty much what I refer to is a bulk activity, whatever is stored gets protected to a greater or lesser degree irrespective of the information content. An email with a lot of technical information that may be the latest thinking on a specific problem or topic authored by a leading specialist will usually get the same protection as an email announcing the Joe Schlitz retirement party. Email is email. Or, it may be even more generalized than that, anything stored on any NT server anywhere in the company gets the same protection. OK - Not always true everywhere, but true enough often enough to make for impossible management of information based on content.

Information has a life cycle. In this respect it is no different than a physical asset. And like physical assets, information generally depreciates with time. There is no such thing as a collectors market for aged information. I’m not talking about rare documents, I’m talking about the intellectual content. Anyway, not much important information is retained on physical documents at this time, it’s pretty much all digital these days.

Storing information is not free. There’s the cost of the physical devices needed for storage. There’s the additional cost of whatever security measures are applied to it. There’s insurance, etc. etc. But the biggest cost is the potential legal liability that can arise from information retention which could have been legally destroyed. In a litigation situation this stale information which long ago outlived its productive use may be the subject of discovery. Destruction of this information at that point is prohibited as it may be considered evidence. Even seemingly innocent information like a calendar or meeting minutes or the list of folks that showed up a Joe’s retirement party may be used to prove someone was in attendance at a certain time and place which might form the basis of case. I’m not at liberty to provide details, but I know of specific situations in which retained notes and memos that should have been destroyed ended up as instrumental evidence costing the company millions of dollars.

The reason a valuation is not placed on information is simple. It does not get reflected as an asset on the balance sheet, hence, no matter what it is, it has no material value (OK, patents, trademarks and copyrights are somewhat different. But they reflect only a miniscule amount of proprietary information held by most companies).

I’m not advocating revising GAAP to include information as an asset. But, companies often keep book on things that do not show up on quarterly reports - or anywhere for that matter. I did put together some proposals (I’ll spare the readers of this board) on information valuation and associated depreciation. The purpose was to get folks to recognise the information life cycle and to develop archive and purge policies for information. All of this was pretty much roundly ignored. The company just kept buying more storage capacity every year as more information accumulated. When the information was retained on physical media, there was at least recognition of the physical space required to house it. Now that it’s all somewhere remote on the network or in the cloud, it seems to have no physical presence whatsoever.

The value of R&D is in the information produced, not the amount of money expended. Tinker pointed out that a lot of R&D dollars end up producing nothing of utility. Nevertheless, from an accounting point of view, it has exactly the same value as the nuggets that lead to innovation and breakthrough.

I’m not sure this is relevant to the topic which was valuation of a company with products that do not require additional plant in order to produce additional revenue. My point is that traditional accounting rules and policies are not well suited to companies that make money by producing and selling information. In a way, we’re stuck. Really, all we have is revenue and earnings and all the traditional GAAP categories and rules. Is any one piece of ERP s/w better than another? At this point one would be forced to concede that it’s pretty much a commodity item. One might decide to go with Oracle rather than SAP or what have you, but functionality won’t be the deciding factor in most cases.

Saul primarily looks at the business model and growth as measured by revenue, earnings, customers, etc. as well as other numbers provided by traditional financial reports and makes decisions, often while admitting he does not have much understanding of the specifics of what the companies products actually do.

He’s had a lot of success. Maybe what we got is good enough for the purpose of making investments.

3 Likes

A lot of this can be put into a good “other” category called “goodwill”. I am not a tech company or anything, but I retain clients based upon my years in the business, my maintenance of my website, and the intellectual property that I have drafted up over the years (which apparently outnumbers and is better in quality than what my competitors put up by paying outsourced writers to do).

This should give my firm a higher multiple, just as the brand of Coke gives Coke a higher multiple as there is so much goodwill value put into that brand name. It is intangible but as real as any other form of value in a business.

Perhaps, instead of getting into this thicket, we can just figure out how much goodwill value this R&D, branding, and other intellectual property has, that is not put on the balance sheet, and add it into goodwill, to help demonstrate why a premium valuation is justified.

For Microsoft, the goodwill value with Windows has to be worth, well you figure out how many billions and billion and billions of dollars. CUDA may have similar value for Nvidia, EOS for Arista, SHOP certainly has goodwill, perhaps more goodwill value than anything else on their balance sheet. In fact SHOP probably has a larger CAP than Nvidia, or Arista, as an example, because it is difficult to think of any disruption that would take them out.

Perhaps a simpler method to handle the problem. Better use goodwill.

Tinker

1 Like

…or is it in the technology to build a motor, a transmission, a suspension system, the styling of the car, the seats, the handling, and increasingly, the software that drives all of the above. I would argue is in the technology of not only the cars they sell, the technology of how to bring 5,000 or more different parts together into a car I, you and everyone else wants to buy.

You are making a common mistake. Your statement assert that R&D only applies to product development. I can assure you from inside experience that a ton of money goes into manufacturing R&D. That’s money spent on the technology required to build and test products rather than product development.

We don’t build TVs in this country anymore. That’s not because we can’t design the product, rather we no longer have the technology to build flat panel displays. We know what to do, but we don’t know how to do it more efficiently than the Chinese, Taiwanese or Koreans. The cost differential is not in the labor, it’s in the manufacturing technology.

As for Boeing’s use of program accounting, well it’s not just Boeing, it’s pretty much the entire aerospace and defense products industry. I’m not sure where else these accounting methods are used, but there are things that make these kinds of products quite different from most other industries.

Boeing spends years in product development. You may have never heard of the 7J7, that was an airplane in development during the 80s. It was intended to serve the market below the 737 in terms of passenger load and routes. Boeing was spending over a million dollars a day (R&D) for over a year before they killed the program as they determined they didn’t really have a market for the plane. What happened to all the R&D spend? Some of it proved useful in the development of the 787.

And there was the SST, the Sonic Cruiser the People Mover (a rail project), the Hydrofoil (boats, they actually made and sold a few), Connexion, a failed airborne internet product (it grew out of an R&D project on antennas). Big companies like Boeing pour a lot of money into R&D that ultimately provides some benefits to the company, but often it is via the contribution to the body of knowledge rather than the specific product it was intended for. Had it not been for Boeing’s contribution to the B1B bomber they would never have built the 787. While the 787 has a ton of product innovation, it was the manufacturing technology that made the plane possible.

Every one of those product R&D programs was accompanied with manufacturing R&D. They don’t wait until the product design is 80% complete before they start figuring out how to build it.

5 Likes

Here’s a related article by Gu and Lev - “Time to Change Your Investment Model” -

https://www.cfapubs.org/doi/pdf/10.2469/faj.v73.n4.4

Summary:

We demonstrate empirically that the gains from predicting corporate earnings, or consensus hits and misses — an activity at the core of most investment methodologies — have been shrinking fast over the past 30 years. We identify the main reasons for this loss of earnings relevance and propose an improved alternative to current investment methodologies, one that focuses on the “strategic assets” of the enterprise and their contribution to maintaining the company’s competitive edge. We demonstrate this investment methodology using subscription-based companies.

1 Like