[HN Gopher] Minsky Moments in Venture Capital
___________________________________________________________________
Minsky Moments in Venture Capital
Author : imartin2k
Score : 112 points
Date : 2022-02-21 10:03 UTC (12 hours ago)
(HTM) web link (pivotal.substack.com)
(TXT) w3m dump (pivotal.substack.com)
| civilized wrote:
| Very interesting.
|
| It seems to me that the basic punchline of Minsky theory is "the
| financial system is designed to encourage momentum investing". Or
| maybe even "everyone is a momentum investor, some just do it with
| extra steps".
|
| Even if you're not consciously a momentum investor, you will use
| or be presented with so-called "risk" metrics which are really
| just looking at past performance, in a way that mostly amounts to
|
| Stonk go up = "low risk"
|
| Stonk go down = "high risk"
|
| If you use such metrics, you're just a momentum investor with
| extra steps.
|
| And if everyone is just a momentum investor with extra steps, our
| markets will of course boom and bust, as a market full of
| momentum investors must do.
| lordnacho wrote:
| One thing I always wondered as a hedge fund guy was where all the
| reports of VC/PE doing badly went. You always hear about some
| hedge fund blowing up or having a terrible month, but I seem to
| never hear any bad news from the private markets world. In fact
| it's all "we returned 10x capital" which is pretty good even if
| it takes a few years, especially if there's no losses. Even
| Softbank seems to be doing okay considering the sheer scale of
| bad press.
|
| Of course PE is a bit different to VC in the leverage, and might
| be a candidate for Minsky moments through the traditional credit
| spiral.
|
| This essay gives a plausible alternative route for VC and
| clarifies what actually makes the spiral happen.
|
| But for both I just struggle to think of many bad events other
| than WeWork, which hasn't even died.
| JumpCrisscross wrote:
| > _I seem to never hear any bad news from the private markets
| world_
|
| The names are less recognisable, so the stories don't get as
| much traction in the popular, paywall-free press. But plenty of
| funds shut down, some quite spectacularly, _e.g._ Rothenberg.
| atdrummond wrote:
| It's mainly confined to industry specific news sources but
| alternative asset data firms like Preqin definitely report on
| this info. And it makes sense that this info would be out there
| for two reasons: 1. Institutional investors, like CPPIB, who
| still invest in alternatives need a way to objectively compare
| funds and 2. VCs and PE, just like HF, both under-perform the
| markets outside a small set of firms that return genuine alpha
| consistently - just like hedge funds.
| MichaelZuo wrote:
| Didn't Uber have a down round? They didn't implode of course
| but it seems like it did significantly affect their trajectory.
|
| I can't think of any big YC company with a down round + death
| spiral, but I think maybe one of the Techstars companies?
| arbuge wrote:
| Homejoy perhaps?
| MichaelZuo wrote:
| Wasn't a unicorn I believe but seems close enough.
| swrj wrote:
| There is a good list of startups that went bankrupt here:
| https://www.cbinsights.com/research/startup-failure-post-
| mor... Katerra seems like the latest big one - $1.5B in
| funding that filed for chapter 11 last year.
| JumpCrisscross wrote:
| > _Without mark-to-market, there's no chance of a risk-reduction
| spiral_
|
| Firms trading in the private markets for peanuts continue being
| marked at last round on funds' books. Which works fine if the
| fund broadly performs and then writes off or distributed those
| stumps at EOL. Until those IRRs fall and LPs begin demanding
| market marking that status quo remains stable.
| hncommenter13 wrote:
| Interesting piece. But speaking as someone who was formerly a
| very junior VC through the dot-com era, there most certainly can
| be a negative spiral.
|
| The public and private markets aren't as distinct as they might
| appear to be. A VC buying shares in a private company at
| valuation X must believe that a sale is possible at a big
| multiple of X, and soon. Some VC will be the last investor before
| the company goes public or is acquired. And that last private
| investor has to sell to another buyer, either a strategic
| acquirer with cash (or highly-valued stock) or an investor making
| a purchase in an IPO. And if those exits don't look as rosy as
| they used to (seen the share price movements of publicly-traded
| tech stocks lately?), the whole thing runs in reverse.
|
| Worse, if the companies needing financing aren't cash-flow
| positive or profitable (and few are), existing investors' stakes
| will be diluted as prices drop. Investors might want to slow the
| pace of investments to reserve cash to fund the needs of their
| existing companies, rather than take bets on additional companies
| needing cash.
|
| Also, while speed is good for startups, "time diversification"
| used to be considered a good thing for VC investors, who really
| are playing a portfolio game. The worst-performing funds from the
| dot-com era were those raised and invested in 2000, just before
| the peak of the bubble. Of course, at the time, no one knew it
| was the peak.
|
| Almost no one working in VC now would remember it, but there was
| a short recession in the early 90s that greatly affected the VC
| industry. The fund I worked for had been founded in the mid-80s,
| and reading the old investor letters was fascinating. Admittedly
| early stage tech was a far smaller industry back then (the
| dollars thrown around now make the deals I worked on in the dot-
| com era look positively quaint), but so were the burn rates.
| JumpCrisscross wrote:
| > _while speed is good for startups, "time diversification"
| used to be considered a good thing for VC investors, who really
| are playing a portfolio game_
|
| To expand on this, as the article notes, the old game might
| have been (stylised) ten Series A investments, three of those
| raise a B and one raises a C. Today, all ten raise a B and then
| three months later a C. Valuations (perception of risk) go up
| (down). But has actual risk been reduced?
|
| _The Information's_ "The End of Venture Capital As We Know It"
| [1] argues that yes, software start-ups _have_ become less
| risky over the past decade. I agree with this in part.
| (Cautiously. I make more money when Silicon Valley valuations
| go up, so of course I'd like that argument. It also sounds like
| "this time is different.") Even if true, we may have overshot
| the mark. In a way, those mis-placed follow-on bets on doomed
| unicorns are VC's analogy to leverage--it's amplifying a single
| company's effects on the portfolio.
|
| [1] https://www.theinformation.com/articles/the-end-of-
| venture-c...
| lumost wrote:
| Objectively, software bets are less risky than 20 years ago.
| There are more well understood business models, more ways to
| reach customers, fewer technical risks (putting a consumer
| website meant dropping 7 figures on hardware and hoping you
| had the right team to make things work), and a better
| understanding of what a defensible business looks like.
|
| On the flip side a reduction in risk is no guarantee of
| success, there are new risks related to having ~100 copy cat
| companies - or having your business replicated by a mega-cap.
| Lower risk means lower barrier to entry.
| cs702 wrote:
| One other thing that happened in the dot-com era, which is also
| happening today, is that many new companies would spend fresh
| funds raised from VCs and even from IPOs _to buy products and
| services from each other_ , spending money aggressively to
| deliver such products and services, and generating revenue
| growth that would look impressive in the short run... but
| ultimately would prove unsustainable. Such growth can last only
| as long as there is an ongoing supply of fresh capital!
|
| At the extreme, some companies in the dot-com era engaged in
| dubious "round-trip revenues" behavior, e.g., agreeing to buy a
| certain dollar amount of another company's products/services
| only if the other company agreed to do the same, with neither
| company actually needing to do so for ordinary business
| purposes. I don't know if this is happening today too, nor to
| what extent.
| onlyrealcuzzo wrote:
| This seems to be A LOT of Series A SaaS.
|
| Theoretically, who cares. If you can convince people to buy
| your product - even if only to scratch each other's backs -
| that's better than all the other people who can't / don't
| have the network.
| jacquesm wrote:
| > Some VC will be the last investor before the company goes
| public or is acquired.
|
| Or goes bust.
| JohnJamesRambo wrote:
| This was a wonderful, easy to understand description of the whole
| process. Bravo!
| marquisdepolis wrote:
| This is a good exposition of a normal dynamic in the public
| markets. I don't think it will be as dramatic in the private
| markets because spirals require liquidity and regular price
| updates neither of which happen in private markets.
|
| That said the idea that once you think you know how to do a
| strategy (whether that's in bond trading or in growth rounds) its
| utility decreases as everyone copies it is very true, and a key
| aspect of how markets get efficient.
| blitzar wrote:
| > I don't think it will be as dramatic in the private markets
|
| They used to say the same thing about synthetic CDOs
| JumpCrisscross wrote:
| > _spirals require liquidity and regular price updates neither
| of which happen in private markets_
|
| They do. They're just slower and more opaque.
|
| If institutional backers stop letting Tiger _et al_ raise ten-
| and eleven-figure funds, growth-stage capital could dry up.
| That means term sheets vanishing, expected funding going away.
| Some firms will lean up and survive. Many won't be able to.
| Those that raise will raise at worse terms. All of which
| affects the prices at which these companies' shares trade on
| the secondary market, which feeds back into the fundraising
| difficulty (for funds and companies alike), with the added
| dimension of employee compensation and morale.
|
| For companies with the massive burn and breakneck growth the
| current environment has encouraged, the snap could come quite
| fast. That hole in investors' portfolios could then extend the
| pain to their more-disciplined peers. As another comment notes,
| illiquid markets melt down all the time. They just do so more
| ambiguously and more dramatically, with the bottom falling out
| as the top deckers keep sunning.
| vmception wrote:
| This is all great and novel looks at improved market efficiency,
| but it is impossible to separate this from the Fed's 8 trillion $
| balance sheet and the continued speed of which it buys stuff
| ($100bn per month) to grow that balance sheet
|
| If the Fed stops giving people $100bn/month in new dollars for
| their bonds and mortgage derivatives, will they still invest in
| new VC funds? Will the VC funds have capital to pump these
| startups so fast?
|
| Turn off the Fed spigot and what? Is the confidence in the
| private market really there?
|
| There just isn't a history to know!
| JumpCrisscross wrote:
| > _will they still invest in new VC funds?_
|
| Given the Fed's announcements have prompted a rotation out of
| fixed income and into public equities, probably. Equities fare
| better in times of inflation. Private equity is still equity.
| Valuations may flag.
|
| > _There just isn't a history to know!_
|
| Yes there is. Venture capital, as an asset class, has been
| through multiple crashes and tight-money regimes. (Also,
| nitpick: the Fed hasn't been buying $100bn/month for some time.
| It's currently in the $10 to 30bn regime and winding down.)
| vmception wrote:
| > It's currently in the $10 to 30bn regime and winding down.
|
| Double checking this, I think all of our numbers are off.
|
| My initial number was very outdated, your amount is just the
| amount of the reduction.
|
| I'm reading January was [to be] $60bn in purchases.
| JumpCrisscross wrote:
| > _I 'm reading January was [to be] $60bn in purchases_
|
| January was $40bn, February is $20bn [1]. (There is a
| January release pointing to $30bn [2] from which I was
| remembering my figures.)
|
| [1] https://www.newyorkfed.org/markets/domestic-market-
| operation...
|
| [2] https://www.newyorkfed.org/markets/opolicy/operating_po
| licy_...
| vmception wrote:
| Thanks, right! And the taper goal is to continue reducing
| this month over month
|
| As well as potentially selling some of the stuff it has
| already bought, but more likely just holding things till
| maturity, or maybe a combination
| evancoop wrote:
| The antidote seems to be taking the long view. Minsky moments in
| finance would have been diminished if CEOs of banks were
| compensated based upon the status of the firm in 10 years or if
| their estates' incentives were aligned with the state of the bank
| in 20 or even 100 years.
|
| VC presumably can structure their compensation paradigms
| similarly. Thus, minimizing the probability of ruin becomes a
| more rational choice.
| lumost wrote:
| Looking at a company like $SNOW with a price to revenue
| multiplier of >100x, the core thesis rings true. Companies are
| being priced as if they have already succeeded beyond the
| expectations of all past companies in the space.
|
| To make SNOW's valuation make sense you would need to assume that
| they can get to a _profit_ of 2.6 Billion /year. This would put
| them squarely in the realm of hot, high growth stocks with a P/E
| of 30x. Put differently, SnowFlake's TAM must be able to support
| ~6 Billion in profit to keep a flat valuation and a P/E ratio of
| 13. Assuming they were able to offer the service while returning
| their full gross margin as profit would imply a revenue of 10
| billion or a little over 16x revenue growth priced into the
| stock.
|
| While SnowFlake is a great company with a great product. Is any
| enterprise software company really worth pricing an optimistic
| 16x revenue growth into?
| thedudeabides5 wrote:
| As a competitor to SnowFlake I talk about this 100x price to
| sales in every meeting
| lumost wrote:
| When I see SaaS multiples like this, I worry that it
| incentivizes consultancy behavior. SaaS is attractive as the
| R&D work gets amortized over many customers and many years.
| Customer acquisition costs are often looked at separately for
| this reason as the initial sales engineering/solutions
| consulting/sales work is done once and then the business
| makes profit on the back-end.
|
| At a 100:1 valuation to revenue multiplier, doing ad-hoc work
| to close deals becomes awfully appealing - likewise
| performing ad-hoc work to keep the revenue coming becomes
| very appealing. Pretty soon, you may find that the business
| has become a consulting organization masquerading as a SaaS
| as the platform isn't very compelling without the broader
| consulting business.
|
| A Consulting business typically maintains a 1.2x multiple on
| EBITDA earnings as it tends not to scale very well due to
| both execution quality challenges, and relatively
| low/squeezable margins.
| onlyrealcuzzo wrote:
| 16x over how many years? A lot of companies are growing 40%+
|
| I don't think it's absurd to think AWS and GCP can keep rates
| like this up for another 4-5 years.
|
| If SNOW could do it for 8.5 years - that's ~16x growth.
|
| Still that seems way too optimistic for me - unless I'm purely
| speculating on greater fools willing to pay more.
| lumost wrote:
| aye - effectively this says that SnowFlake will execute
| perfectly for 4 years, or slow to 40% growth for 8.5 years
| and will maintain/reach a healthy margin. For this, the
| market is willing to price this performance in _today_.
| Meanwhile a dollar placed in the S &P 500 would likely 2x
| over the same time frame.
|
| To see a return on a dollar placed into snowflake we'd need
| to start adding in assumptions like "SnowFlake doubles in
| revenue for the next 7 years reaching a revenue of 76 billion
| dollars per year in 2029, surpassing GCP, rivaling major
| cloud vendors".
| onlyrealcuzzo wrote:
| Why? Google's PE isn't going to drop to 13 with NRIRP and
| its current growth rate.
| lumost wrote:
| Google has multiple high margin business lines which have
| grown rapidly for the last 2 decades. Snowflake operates
| in a specific competitive market, with a yet to be
| determined profit margin.
___________________________________________________________________
(page generated 2022-02-21 23:01 UTC)