Hi there! I go by KG, and I love studying the history of business and investing. I’ll be sharing some notes from one Investor/Shareholder letter per weekday (mostly from my compilations) here.
Today’s notes are on Paul Kedrosky’s 2004 article: The Art of Valuation.
*Note: Trying out a new format here (for this week) — Letter printed in full followed by annotated letter. If you like this format, please let me know by dropping me a line via kevin@12mv2.com or commenting below. Otherwise, I’ll be returning to the old format (just annotated letter) next week.
The Letter
Investment valuation is part science, but mostly art. But just because valuation is inherently subjective doesn’t mean we should throw out all the rules. Consider the case of the Google IPO.
How much is Google really worth? Cynics have a ready answer: Google is worth whatever people will pay. In other words, if investors are willing to accord the online search company a valuation of $32-billion, the mid-point of its proposed pricing, then so be it. That is what Google must be worth.
Admittedly, there is some screwy logic to this approach. Say you come up with a highly-detailed discounted cash flow model for Google, one that projects sales and earnings out ten years into the future. Say that you discount all those profits back to the present with your best guess of what a reasonable interest rate would be. And let’s say you come up with a valuation that implied $50 a share would be about right for Google’s shares.
Does that mean you should bid $50 a share for Google? You could, but you would get, at best, psychic satisfaction from bidding the “right” price, but not any corresponding stock. Google shares will have all sold at some much higher price, perhaps $120 or more. After all, the company has made it clear that it will be selling stock somewhere between $108 and $135. Pretending otherwise – even if pretending otherwise seems perfectly financially rational – is not very sensible.
So, does that mean you should abandon all valuation models and purchase Google at its likely valuation of around $32 billion and the corresponding trailing price-to-earnings multiple of more than 100? Does it mean that Google has abandoned all pretense of valuing a company, perhaps just picking this number based on some arbitrary multiple of Liechtenstein’s GDP?
No. The path that Google and its bankers took to this valuation is almost certainly very similar to the path that venture capital firms take in valuing firms before early-stage investments. Google and its bankers undoubtedly looked at how comparable firms were valued and relied heavily on that in coming up with an estimate for Google’s worth. What companies? Internet stalwarts like Amazon, eBay, and Yahoo would have headed the list.
With the preceding in mind, the following table summarizes the price-earnings ratios for four Internet stalwarts and their valuations based on twelve-month forward (FTM) earnings. You can see that Yahoo leads the list, trading as it does at 89 times next year’s earnings. Ebay is only slightly behind, sitting in at 63 times earnings, and so on down. Where would Google fit in? Based on reasonably defensible estimates, Google’s 2004 earnings could come in around $2.90 a share. If you assume a $122 IPO share price, that would put Google’s forward earnings multiple at 42, right between Amazon.com and Ebay.
Is that so bad? At face value, not really. After all, Google has better margins than Amazon, is growing comparably to Ebay, and is more profitable than Yahoo. It doesn’t seem unreasonable to value Google similarly to how those companies somewhere around where those firms are valued. After all, Google will have grown revenues almost 100% year-over-year from 2003 to 2004, so applying a 42 time earnings multiple to 2004 earnings isn’t completely implausible.
On the other hand, there are some pesky risk factors. For example, none of the above four companies have dual-class share structures like Google does; they don’t, in other words, entrench control in insiders despite having done an IPO. Some sort of discount needs to be applied to Google’s valuation for the share structure. Similarly, discounts should be applied to Google’s price based on the number of insiders and venture firms selling shares in the IPO, and also for the relative newness of the company. (Search engine firms seems to have the lifespan of bottle rockets, with successive generations launching and fizzing out in short order over the last decade.) And there are other discounts worth applying. For example, there is a huge overhang of Google stock waiting to come crashing into the market. The company is only selling 10 percent of its shares; there are, in other words, nine shares waiting to come tumbling into the market for every Google share being sold in the IPO. And then there is the margin issue, with Microsoft and Yahoo poised to cause trouble in both the search and ad placement markets, thus making it unlikely that Google will be able to sustain its lofty current margins.
Where do we end up? At some number lower than the 42-times-earnings figure I suggested above. How much lower is impossible to say, but that is beside the point. Valuing Google in comparison to its peers is useful, but only insofar as you take into account both the things that make the company similar to its peers, and the things that make it different. Perhaps the real value in looking at Google in these terms is that we are now able to say that we now know why Google is valued the way it is valued – it is being valued based on how its peers are valued – and we can make a decision to invest, or not invest, on that basis. Because relative valuations are not always correct. They are, in a sense, right until they are wrong. As the following figure shows, relative values oscillate around some norms, but they can also drift far from those norms. For example, price/sales multiples in the public markets have been almost everywhere except their long run average of about 1.7. That said, woe betide anyone who tries to use bubble era relative valuations in the current market, or vice-versa. You would have no takers, so it’s useless pretending that what everyone else thinks doesn’t matter.
In a sense, relative valuations represent the wisdom of crowds, to steal the title of New Yorker writer James Surowiecki’s recent book. We can be as clever as time allows in coming up with a valuation for a company based on their intrinsic financials, but it doesn’t really matter. The masses in the markets, whether it is public markets, as in the case of Google, or private markets, as in the case of most venture investments, will make their own decisions. As investors it is up to us to decide the basis for the valuations being accorded firms, and then to decide whether we want to play in markets where we don’t like the valuations. Because if we do play in expensive or seemingly mispriced markets, like the Google IPO, we are going to have to protect ourselves from those inevitable times when the crowds aren’t being very wise at all.
Annotated Letter
Investment valuation is part science, but mostly art. But just because valuation is inherently subjective doesn’t mean we should throw out all the rules. Consider the case of the Google IPO.
Even art has rules — You may say, “Well, Picasso was known for breaking all the rules.'“ But did you know he was also classically trained from a young age?
Focus on first principles — throw everything else out
Classic Value Investing Quote — “Valuation is an art, not a science”
How much is Google really worth? Cynics have a ready answer: Google is worth whatever people will pay. In other words, if investors are willing to accord the online search company a valuation of $32-billion, the mid-point of its proposed pricing, then so be it. That is what Google must be worth.
Publilius Syrus: “Something is only worth what someone is willing to pay for it.”
Buffett: “Price is what you pay. Value is what you get.”
Paul seems to be a value investor
Admittedly, there is some screwy logic to this approach. Say you come up with a highly-detailed discounted cash flow model for Google, one that projects sales and earnings out ten years into the future. Say that you discount all those profits back to the present with your best guess of what a reasonable interest rate would be. And let’s say you come up with a valuation that implied $50 a share would be about right for Google’s shares.
“Screwy logic” — lol I love his writing
10 year DCF never made sense to me — I recall some Philip Tetlock research that said even the best forecasters can only predict 1-2 years ahead, and they’re a rare bunch. Most people do no better than a monkey throwing darts (but at the same time, they say give monkeys typewriters and enough time and they’ll write Shakespeare — personally I think it’d be a miracle if a monkey would actually pick up and throw a dart and actually hit the dartboard)
Personally, I think DCFs are good for thought experiments and okay to use to as tools and support, but really they’re just so easily manipulated. Some of my favorite sayings about models:
“When in doubt, MSU (Make Sh*t Up)”
“Garbage in, garbage out”
Does that mean you should bid $50 a share for Google? You could, but you would get, at best, psychic satisfaction from bidding the “right” price, but not any corresponding stock. Google shares will have all sold at some much higher price, perhaps $120 or more. After all, the company has made it clear that it will be selling stock somewhere between $108 and $135. Pretending otherwise – even if pretending otherwise seems perfectly financially rational – is not very sensible.
Even if my valuation is $50 a share (with conservative assumptions), I wouldn’t pay that much. I’d want a margin of safety or see solid growth potential
TBH I was about to say I wouldn’t use a DCF for Google, but at the time, they didn’t exactly have comps either — and very few, if anyone, could have predicted the rate at which they grew. Even if there were, comps are just another form of DCF analysis, and you got to do something.
It’s quite clear why value investors have been mostly crushed over the past 15-20 years (my hypothesis is the digitization of the world has rendered traditional economics moot as scarcity has mostly disappeared (see Albert Wenger’s book World After Capitalfor some more on this))
You need to be intellectually honest — pretending will get you nowhere, especially if the pretending “makes sense”
So, does that mean you should abandon all valuation models and purchase Google at its likely valuation of around $32 billion and the corresponding trailing price-to-earnings multiple of more than 100? Does it mean that Google has abandoned all pretense of valuing a company, perhaps just picking this number based on some arbitrary multiple of Liechtenstein’s GDP?
Lol. It’s sad because it seems like so many venture capitalists do this today. Let’s slap a random multiple on this.
Remember when Bill Gurley wrote about the “Keys to the 10x Revenue Club”? When a 10x revenue multiple was supposed to represent a supremely moated company? Well VCs nowadays are paying 50x, 100x revenue multiples for startups with no moat whatsoever.
Reminds me of a funny story when a friend had just started at a PE shop’s new growth practice. They were pitching a company to the Investment Committee, a company with a revenue multiple ~6x. To the typical early/growth investor, a ~6x multiple is a red flag — there’s something wrong with the company. Well, one of the IC members stopped them there and said, “6x revenue?!? We don’t even pay 6x EBITDA for our LBO deals, and this company doesn’t even have EBITDA!!” Needless to say, they did not invest, and my friend knew he was in for a rough time.
I have friends in venture who can’t even tell me how to calculate free cash flow, much less build out an entire DCF (but then again, I have some friends at hedge funds who can’t either lol).
In case you’re wondering, the *basic* equation for Free Cash Flow is: EBIT(1 - Tax) + non-cash expenses like Depreciation and Amortization +/- Change in Net Working Capital - Capital Expenditures
No. The path that Google and its bankers took to this valuation is almost certainly very similar to the path that venture capital firms take in valuing firms before early-stage investments. Google and its bankers undoubtedly looked at how comparable firms were valued and relied heavily on that in coming up with an estimate for Google’s worth. What companies? Internet stalwarts like Amazon, eBay, and Yahoo would have headed the list.
Comps
Public investors keep shifting to private multiples, and private multiples just keep growing and growing and growing
Interesting parallel/comp would be golf club lofts: The Evolution of Lofts. Over the years, club manufactures lower the loft on clubs in order to say “they now hit further”
With the preceding in mind, the following table summarizes the price-earnings ratios for four Internet stalwarts and their valuations based on twelve-month forward (FTM) earnings. You can see that Yahoo leads the list, trading as it does at 89 times next year’s earnings. Ebay is only slightly behind, sitting in at 63 times earnings, and so on down. Where would Google fit in? Based on reasonably defensible estimates, Google’s 2004 earnings could come in around $2.90 a share. If you assume a $122 IPO share price, that would put Google’s forward earnings multiple at 42, right between Amazon.com and Ebay.
Is that so bad? At face value, not really. After all, Google has better margins than Amazon, is growing comparably to Ebay, and is more profitable than Yahoo. It doesn’t seem unreasonable to value Google similarly to how those companies somewhere around where those firms are valued. After all, Google will have grown revenues almost 100% year-over-year from 2003 to 2004, so applying a 42 time earnings multiple to 2004 earnings isn’t completely implausible.
Based on comps, if anything, Google should probably be trading higher
On the other hand, there are some pesky risk factors. For example, none of the above four companies have dual-class share structures like Google does; they don’t, in other words, entrench control in insiders despite having done an IPO. Some sort of discount needs to be applied to Google’s valuation for the share structure. Similarly, discounts should be applied to Google’s price based on the number of insiders and venture firms selling shares in the IPO, and also for the relative newness of the company. (Search engine firms seems to have the lifespan of bottle rockets, with successive generations launching and fizzing out in short order over the last decade.) And there are other discounts worth applying. For example, there is a huge overhang of Google stock waiting to come crashing into the market. The company is only selling 10 percent of its shares; there are, in other words, nine shares waiting to come tumbling into the market for every Google share being sold in the IPO. And then there is the margin issue, with Microsoft and Yahoo poised to cause trouble in both the search and ad placement markets, thus making it unlikely that Google will be able to sustain its lofty current margins.
VCs are more partial to dual-class structures. It gives the visionary founders more control and the ability to think and execute for the long term. I’d argue today this is a plus rather than a minus (although it seems with Palantir’s going public people are criticizing multiple-class structures again — although this is probably more so just because it’s Peter).
Discounts should also be applied based on insiders and VCs selling (this I don’t really get why tbh) and relative newness (makes sense, unproven companies should trade at discounts — although how does this change given companies are staying private longer?)
Lots of discounts worth applying (spoken like a true value/public markets investor):
Stock overhang — when shares flood the market and scarcity/demand goes down, so will the stock price
Margins — basically, “Your margin is my opportunity” (Bezos)
Where do we end up? At some number lower than the 42-times-earnings figure I suggested above. How much lower is impossible to say, but that is beside the point. Valuing Google in comparison to its peers is useful, but only insofar as you take into account both the things that make the company similar to its peers, and the things that make it different. Perhaps the real value in looking at Google in these terms is that we are now able to say that we now know why Google is valued the way it is valued – it is being valued based on how its peers are valued – and we can make a decision to invest, or not invest, on that basis. Because relative valuations are not always correct. They are, in a sense, right until they are wrong. As the following figure shows, relative values oscillate around some norms, but they can also drift far from those norms. For example, price/sales multiples in the public markets have been almost everywhere except their long run average of about 1.7. That said, woe betide anyone who tries to use bubble era relative valuations in the current market, or vice-versa. You would have no takers, so it’s useless pretending that what everyone else thinks doesn’t matter.
Understand the lens in which you are looking — and make a decision based on that (aka understand your biases)
Valuations are, in a sense, right until proven wrong (reminds me of the old saying — “never wrong, just early” although others point out the “early is the same as wrong.” How does timing play in?
Keynes: “The Market can remain irrational longer than you can remain solvent” — mean reversion may happen, but not before you go bust
Understand where in a cycle you are — if you’re in a bull market, use bull market valuations. If you’re in a bear market, use bear market valuations. It makes no sense to use bear market valuations in a bull market or bull market valuations in a bear market.
The market matters — at the end of the day, the market is the intersection of buyers and sellers. What other people think definitely does matter. If there’s enough “wrong people” they’re right.
In a sense, relative valuations represent the wisdom of crowds, to steal the title of New Yorker writer James Surowiecki’s recent book. We can be as clever as time allows in coming up with a valuation for a company based on their intrinsic financials, but it doesn’t really matter. The masses in the markets, whether it is public markets, as in the case of Google, or private markets, as in the case of most venture investments, will make their own decisions. As investors it is up to us to decide the basis for the valuations being accorded firms, and then to decide whether we want to play in markets where we don’t like the valuations. Because if we do play in expensive or seemingly mispriced markets, like the Google IPO, we are going to have to protect ourselves from those inevitable times when the crowds aren’t being very wise at all.
The market is a weighing machine — it will make its own decision
Fish where the fish are (best if the fish are in a barrell and there are no other fishermen) — Munger has been hawking investing in China (he does it through Lu Li) for the past 5-10 years, even though no one listens
Buffet: “Rising tide lifts all boats” — if you play in dangerous markets, be prepared to capsize and drown when they inevitably correct
Wrap-up
If you’ve got any thoughts, questions, or feedback, please drop me a line - I would love to chat! You can find me on twitter at @kevg1412 or my email at kevin@12mv2.com.
Please DM or email me any time — to share non-obvious intel, views and correct or solicit mine. I appreciate your continued support and partnership :D.
As many of you know, making compilations and this newsletter is a (very time consuming) passion project for me. Multiple people have reached out about how they can support me so that I continue, so I set up with a few ways:
I prefer this format too. I love your annotated letter tbh!
I prefer this format! This is great.