Minsky Moments in Venture Capital(pivotal.substack.com)
pivotal.substack.com
Minsky Moments in Venture Capital
https://pivotal.substack.com/p/minsky-moments-in-venture-capital
35 comments
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.
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.
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.
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.
> 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...
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...
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.
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.
> Some VC will be the last investor before the company goes public or is acquired.
Or goes bust.
Or goes bust.
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.
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.
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?
I can’t think of any big YC company with a down round + death spiral, but I think maybe one of the Techstars companies?
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.
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.
> 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.
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.
Possibly silly question: how does one transition from software engineering to hedge fund management?
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?
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?
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.
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.
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".
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".
Why? Google's PE isn't going to drop to 13 with NRIRP and its current growth rate.
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.
I mean - I agree. But who's to say people aren't willing to bet this as yet to be determined business ends up being the best business of them all?
I mean there will always be a next rocket ship, and it’s obviously clear that people are willing to take this bet.
However at the same time, I’d be curious why a lot of money is taking snow as being the next google as a given. My bet is because other institutional investors are taking this as a given, and the time value of capital has been distorted by low interest rates.
If this illusion pops before snow hits 10 billion in revenue it’s likely going to cost investors dearly. As of right now a plausible downside of 95% exists should the market choose to take the common valuations of 2016.
However at the same time, I’d be curious why a lot of money is taking snow as being the next google as a given. My bet is because other institutional investors are taking this as a given, and the time value of capital has been distorted by low interest rates.
If this illusion pops before snow hits 10 billion in revenue it’s likely going to cost investors dearly. As of right now a plausible downside of 95% exists should the market choose to take the common valuations of 2016.
As a competitor to SnowFlake I talk about this 100x price to sales in every meeting
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.
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.
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!
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!
> 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.)
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.)
> 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.
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.
> 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_policy_...
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_policy_...
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
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
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.
VC presumably can structure their compensation paradigms similarly. Thus, minimizing the probability of ruin becomes a more rational choice.
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.
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.
> 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.
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.
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.
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.
> 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.
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.
> I don't think it will be as dramatic in the private markets
They used to say the same thing about synthetic CDOs
They used to say the same thing about synthetic CDOs
This was a wonderful, easy to understand description of the whole process. Bravo!
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.