Hacker News Reader

Comment by tptacek | Hacker News Reader

Parroting Andreesen from Twitter last night:

The valuation number is misleading. You're looking at the snapshot price of a relative small fraction of Airbnb's equity at a particular moment in time; just as 1,000,000 shares of XYZ will likely trade at a lower price than 1 share, so too might the valuation be lower for the whole company.

More importantly, these aren't common shares; they're shares with privileges attached to them, and thus command a higher value.

Finally, VC's perspective towards share valuation is different than that of a "value investor"; a VC looks at the shares as if they were options, with a limited (1x investment) downside but an uncapped upside.

These aren't my insights; they're taken from a bulleted list Andreesen twerped last night. They sound convincing to me.

View on HN

This 'new valuation debunking' can also be overdone. There's never one single "valuation" - everything is dynamic, contingent on assumptions, and subject to surprising discontinuities.

The naive extrapolated "post-money valuation" number is still significant under many scenarios, and perhaps even the 'modal' (most-likely) or median scenario envisioned by the business insiders who are selling the shares.

After all, it's the exact price that, assuming a 1X liquidation preference, the venture would have to appreciate beyond for the VC to make any positive return – which is after all the whole point. It's the exit value at which the insiders take exactly the negotiated dilution, not extra dilution to protect the investors. (If the insiders' own reserve value was less, they'd likely be selling larger stakes, or the whole company.)

Generally, also, minority stakes like this have a discount to the full-company value because they're at the mercy of the majority/control stakes. (Though, preferences affecting governance also affect this.)

At such a large valuation, and such a small (5%) stake, I wonder what preferences were really offered. It looks like a raise from strength, like Facebook's pre-IPO raises – another time when armchair skeptics scoffed at the implied valuations (which in hindsight were bargains). Maybe the preferences aren't that big in such a case.

Usually, you pay a premium for large equity stakes, because they come with more practical control. (Witness every tender offer takeover ever.) The other two points are more convincing, however.

I don't think pmarca/tptacek mean the price would be lower if someone were buying 1,000,000 shares all at once. Rather, if AirBNB tried to sell 1 share, it would naturally go to the person willing to bid the most for it, and that price would be higher than the average selling price if 1,000,000 shares were released.

I think you're stating their point accurately. But what I mean is that in the usual case, the value of 1,000,000 shares is actually greater than 1,000,000 x (price of one share), because you're getting something more with those million shares: control. That said, there may be countervailing forces at work in tightly-held startups that don't apply to liquid capital markets.

Let me try to spell this out in a bit more detail. I think you know everything I'm about to say, but I want to make sure I understand it:

You'd be crazy to pay more than a few cents more than the market price for 1 share of AAPL, because that price represents the value the most marginal AAPL investor places on the stock: Offer him a penny more and he'll sell it to you. But there are investors standing behind him who think AAPL is worth more than the current price. As you need to buy out less marginal investors, the price to get them to hand over their AAPL shares will rise. Thus 1,000,000 shares will cost you more on average than one regardless of the value of control.

The same works in reverse with AirBNB: If they sell one share, it will go to the most enthusiastic investor who thinks they're a future $1T company. If they try to sell 1,000,000, they have to work their way down into those investors who believe AirBNB is a more marginal opportunity, and the average price will decline.

This is right, but isamuel is also right in that there is the separate effect of the price of control, such that in reality it's harder to say what would happen in the AirBnB case because of the interaction effects.

It would be interesting to probe into further, but the most directly applicable data is probably fairly hard to get ahold of, given that it's investment/acquisition offers not taken in lieu of the other.

It works both ways. If stock currently has 10m shares in circulation, selling 1,000,000 shares you expect to get significantly lower price per share, unless you do it very slowly on a backdrop of high demand for shares. On the other hand, if you want to buy a million shares you can expect to pay significantly higher price compared to buying one share for exactly same reason - price depends on supply vs demand.

In one case of selling shares you are inflating supply and thus dropping price, and when you are buying shares you are inflating demand thus raising the price.

OTOH, if an investor bought 0.5M shares at $x, and then said "but I want MORE shares", AirBnB could reasonably insist on a higher valuation for the next 0.5M, since the buy-side excitement is so high.

Supply vs Demand is tricky....

#1-13 https://storify.com/MDanijel/marc-andreessen-on-private-valu...

Speaking of Twitter last night:

https://storify.com/ramcha/danielle-s-truthstorm-tips-on-sta...

Extreme example that illustrates the point fantastically: 37signals valued at $100 billion after investor buys 0.000000001% of the company for $1. http://signalvnoise.com/posts/1941-press-release-37signals-v...

That really doesn't illustrate anything besides silliness.

It does illustrate how these don't mean anyone actually has $10 billion cash[1], but that wasn't the point here.

[1]: which does still confuse many as shown and explained here: http://money.stackexchange.com/questions/11368/how-are-start...

The problem with this is that they set themselves up for a down round. The smart money will clearly come in at no more than $50 billion :).

The valuation is misleading for what purpose? Pre-IPO valuations are usually quoted as an estimate of post-IPO value.

The difference between the private and public equity markets is scarcity of stock. Liquid dollars from wealthy investors chasing a small supply of transferable stock plumps valuations. Large rounds mitigate the risk that the company is selecting obscene price points.

> a VC's perspective towards share valuation is different than that of a "value investor"; a VC looks at the shares as if they were options

Value investors don't invest in IPOs - growth investors do. IPOs are marketed and bought on the basis of their asymmetric upside potential. VCs have no exclusive claim to optionality (originally a PE term).

> these aren't common shares; they're shares with privileges attached to them

VC preferred shares generally convert to common upon IPO. Valuations which make sense for privileged shares must thus make sense for the common past the IPO.

This situation reminds me of Microsoft buying a small portion of Facebook at a large valuation. In that situation, it was more about the relationship such a deal would create rather than the actual numbers behind it. By creating a large valuation with a small deal, everyone wins. The investor doesn't have to put a large amount behind it while the startup gets the glory of an impactful validation.

http://www.nytimes.com/2007/10/25/technology/25facebook.html...

All that said, I don't think the price x qty numbers are meaningless.

Marc is far and away the biggest bubble denier out there. When this bubble pops, and it surely will, it will be satisfying to watch him eat crow.

Gather round: here we have a rare example of the North American Ivory-Billed True Ad-Hominem Fallacy in its natural habitat.

It is obviously possible for Andreeson to simultaneously be chief among "bubble deniers" and also correct about private company valuations.

Glad you've taken an introductory critical thinking course. This isn't ad-hominem. Raising doubts of the validity of someone's statement based on inherent conflict of interest is a valid line of argument.

No, it would be a valid argument if the argument was "Marc Andreesen says this, therefore it must be true". It happens that Andreesen said these things, but I pointed that out because I wouldn't want people to think they were my insights.

You'll have to actually address the points themselves, sorry.

> Gather round: here we have a rare example of the North American Ivory-Billed True Ad-Hominem Fallacy in its natural habitat.

This is definitely ad-hominem, yourself.

> It is obviously possible for Andreeson to simultaneously be chief among "bubble deniers" and also correct about private company valuations.

Whatever, dude. Marc's point may be relevant for small slices of small firms, but no one has ever spent $500M on what are effectively equity call options outside of a bubble. As such, I posit in the case of AirBnB his logic is not applicable. So he's only just a bubble denier and not correct about private company valuations.

"no one has ever spent $500M on what are effectively equity call options outside of a bubble. As such, I posit in the case of AirBnB his logic is not applicable."

No one has ever seen a black swan outside of their dreams. If you see a black swan, you must be dreaming.

No, seriously. How is this not a case of "It's never happened before, so it can't be right?" What's the actual counter-argument?

That's some witch-hunt logic there, man..

What does "witch-hunt logic" even mean?

LMGTFY...

http://en.wikipedia.org/wiki/Witch-hunt#Metaphorical_usage

What do you expect him to say? "All these companies I have invested in aren't worth as much as everyone says they are"? "You guys shouldn't be investing in VC (including us) right now"?

The fact is that Andreesen stands to benefit from claiming this is totally normally and rational.

That means there's a chance of personal bias in his opinions, not that he's automatically wrong.

Reply on news.ycombinator.com