How will this bubble end?

I've known for a while that sooner or later I would like to write about "bubble". With Snapchat's rumored $500M fund-raising at $16-19B, I guess that time is now.

First let's get one thing straight: all finance researches show that just like most other things in finance, bubbles are very difficult to call out, if not entirely impossible. Most bubbles only look like bubbles after they bursted.

One can also apply Heisenberg's Uncertainty Principle to the stock market (and any asset market) in the spirit that the observation changes the behaviors of the observed. For example, even the most rigorous and painstaking DCF valuation model needs to decide on a discount rate (or multiple rates for different periods of the firm's life). Such a discount rate is related to the risk (volatility) of the free cash flows in each year of projection. If the risk is low, the discount rate is low and vice versa.

But risk is very difficult to measure even for the least sexy financial products such as bonds, as the recent financial crisis proved to us. It's even more difficult to assess for a disruptive startup trying to change the world — even the great Peter Thiel would be lying to say that he had known the proper discount rates for the years to come when he became the first external investor to Facebook in the summer of 2004, when he put in $500k on Mark in return for 10.2% of the shares, essentially valuing Facebook at $4.9M.

It's not difficult then to visualize that the Heisenburg's Uncertainty Principle works this way in startup valuations:

  • As collectively more money are put into startups, the user indices (or whatever KPIs) grow faster than ever, making the investors and the entrepreneurs more confident
  • As people grow more confident, they perceive the risks to be lower
  • As the perceived risks are lower, they're willing to use lower discount rates in their DCF models
  • As the discount rates get lower and lower, valuations naturally get higher and higher
  • (Then something might or might not happen)
  • All of a sudden sentiments switched into doubtful mode
  • People start to use higher discount rates, leading to lower valuations, and this applies to both the public and the private markets — remember that most startups remain unprofitable accounting-wise even years after their IPOs
  • The observers (investors, analysts) start to change the behaviors (valuations) of the observed (startups, young tech firms)
  • The changed behaviors confirmed the negative sentiments of the observers
  • Both public and private valuations go down the toilet and people start to call out the previous 2 years (or whatever period that's good for story-telling) as a sure-thing bubble

This is nothing fancy but just another equilibrium in human being's daily life, where inputs and outputs are dynamic and affect each other.

A more interesting question would be: if this is a bubble, how would it end?

Note that there's a very different dynamic going on this time compared to the dot-com bubble in the 90's — today VCs are investing late into the rounds of successful startups and delaying the IPOs, usually at jaw-dropping valuations.

Consider Snapchat, a company that looks even flimsier than Twitter when the later hit IPO in November 2013 in terms of potential for monetization. It's now rumored to be valued at $16-19B. To put this into perspective, the venerable BOX that counts all Fortune 500 companies as its clients just went IPO in January 2015. Its current market cap? A bit north of $2.1B.

Obviously sentiments should affect both the public and the private markets, so it looks highly incredulous that a revenue-generating B2B startup that seems to own the future of enterprise softwares is valued by the public market at only 1/8 the price of a weird startup that has shown very little revenue-generating ability so far.

If there's a bubble, it's more likely to be a private bubble than a public one — one might be tempted to say this.

But while we all know how a public bubble usually ends — falling share prices, bankrupt citizens and companies, rising unemployment rates — how would a private bubble end?

Specifically, how would this bubble end?

Venture Capital firms manage money for institutional investors such as pension funds, college endowment funds, etc. If these investments go south en masse, institutional investors will get hurt and indirectly the real beneficiaries behind them, e.g. current and future pensioners for pension funds, college operations for endowment funds, etc.

To the first degree there doesn't seem to be a lot of people going broke directly. Sans any spillover effect where the burst of a private bubble induces the public market to crash as well, it seems that the impact of private bubble bursting will be spread out temporally as such investments are intrinsically long-term. Having 3~4 vintage years of VC investments crashing in private markets does not have the same immediate impact of a public market crash.

Obviously many startup employees would lose their jobs. However that's a much smaller group compared to the people affected when the public market crashes. Also startup employees are already used to companies going bankrupt — that's the nature of business in startups. Even if 100 UI designers will have trouble in the short term to land a job, it's not as if they were like unionized teachers in France who have not thought about this possible outcome.

Given this seemingly lukewarm doomsday, dare we say that a private bubble actually a more merciful bubble than a public one as the downside for the society seems to be borne by the more educated and sophisticated?

I would love to say yes but I fear that it's actually a no. My best guess would be that even if this bubble is largely private, it will spill over to the public market regardlessly, leading to a familiar systematic crash that a pure public bubble would have created when it bursts.

To some degree we have seen this before when Credit Default Swap, a highly private market, brought down Lehman Brothers and crashed the whole public market. However, CDS in 2007-08 was notorious for a trading volume that's many times over the underlying assets, plus lots of leverage in the system. Compared to that, startup investments at worst are putting money in new products that will have poor or no returns. Startup employees still get their salaries. All suppliers still get their checks. All paying customers pay their fees. All non-paying customers benefit from the free services. The dynamic of this private bubble bursting will definitely be very different from the last financial crisis.

It would be really really interesting, from a historian viewpoint, to see how this one ends.


Quick thoughts on Pebble Time and Kickstarter as a retail channel

Life expectancy of status quo