The Inevitable Briefness of Alpha
When I started my career as a bond trader at a quant hedge fund, arbitrage opportunities were relatively plentiful. Not many investors had the knowledge or infrastructure to effectively exploit these opportunities, so the early pioneers in the market made good money.
Of course, it didn’t last. Knowledge spreads, technology diffuses, and arbitrages disappear. Markets asymptote towards ever-greater efficiency. This process is extremely well known in capital markets, and even has a name: ‘alpha decay’.
The Minsky Boom
This is the Minsky boom. Money entering a market boosts returns and reduces volatility, leading to very strong (realized) performance. This attracts more money, which improves performance even more. A positive feedback loop ensues.
And this is perfectly legit! Economies can and do reallocate resources all the time. This is how it works; this is how it’s expected to work.
What Goes Up …
The problem with feedback loops is that they tend to overshoot. Minsky booms in an asset class attract a constant influx of new money, but they also need that influx to continue marking up the price and marking down the risk of the asset class. And as the wise man said, “If something cannot go on forever, it won’t”.
Eventually — and fortunes have been made and lost, trying to predict just when that ‘eventually’ comes — something happens. It could be an exogenous shock like COVID, or a tightening Fed, or an election, or a war; it could be an industry-internal event like a particular firm blowing up or winding down. The new money stops, or maybe just slows down a touch, and prices soften. And that triggers all sorts of nasty consequences.
… Must Come Down
The first thing that happens is that risk-management dashboards start flashing red. Even a slight selloff causes returns to drop and risk to rise; portfolios that seemed well-balanced now appear just a little too aggressive.
Over-extended investors begin to trim their positions. Unfortunately, this causes more price declines, and more volatility, triggering another round of position-trimming. Even conservative investors suddenly realize that their portfolios are riskier than they thought. They sell as well. The ensuing vicious cycle is sometimes called a risk spiral.
Risk spirals are often accompanied by margin spirals. Banks and brokers require leveraged investors to post margins that are proportional to their portfolio risk. As their portfolios become riskier — and remember, the portfolios themselves are often unchanged; all that has changed is the level of volatility in the market — leveraged investors face margin calls. They have to sell assets to service these calls, creating further volatility and downward price pressure; the margin call becomes a self-fulfilling prophecy.
(That’s why it’s always best to be the first firm to unwind positions or call margin. Goldman — disclosure: Matt Levine used to work there — was very good at this, Lehman less so.)
The final domino is a redemption spiral. Seeing declining performance and increasing volatility, LPs in a fund request their money back. To service these redemption calls, the fund has to sell even more of its positions, triggering yet another feedback loop. Boom quickly turns to bust.
Lehman Brothers
This is exactly what happened in credit markets in 2007-08. An influx of cash on the way up, accompanied by the sense that it was a can’t-lose trade; everyone was making money on housing and mortgages and credit. And then an equally unstoppable wave on the way down, as value-at-risk and margin calls and redemption spirals all worked to pull capital out of the credit market.
And Round And Round We Go
In good times, people are encouraged by past success to place larger and larger bets, thinking they're less risky than they actually are, when often it's the very existence of these large bets that drives present success and depresses perceived risk.
Investors believe the trend will continue indefinitely, and become complacent. They invest in lower quality instances of the asset, while increasing their leverage.
And then the music stops. Markdowns lead to deleveraging which lead to more markdowns; the positive feedback loop now operates in the other direction. Eventually, the asset class overshoots as investors become overly risk averse, setting the stage for the next bull market. Stability breeds instability, and vice versa.
Now, all of this is well known. Hyman Minsky fell out of fashion in the 80s and 90s, but his work was rediscovered and widely shared just in time for the GFC. Today it's part of the toolkit for most macro (and many micro) investors; you can also see it in regulatory ideas like market circuit-breakers and systemic backstops.
Is Venture Immune?
But credit is a foreign country; they do things differently there. Let’s talk about early stage tech and venture investing.
At first glance, venture capital seems an unlikely candidate for Minsky dynamics to take hold. Consider:
VCs funds don’t use leverage
They don’t offer redemptions or early liquidity to investors
There are no counter-parties and no margin calls
Volatility is actually good for most VC portfolios (long basket of options)
Without mark-to-market, there’s no chance of a risk-reduction spiral. Without leverage and counter-parties, there’s no chance of a margin spiral. And without investor liquidity, there’s no chance of a redemption spiral. What mechanism could force the liquidation of a venture portfolio, or incept a Minsky bust? For that matter, what’s the mechanism for a Minsky boom in venture?
Like all Minsky booms, there are some genuine truths underlying the dynamics of the venture market today.
Startups are marked up faster than ever — but startups are also growing faster than ever. 3x year-over-year used to be considered strong; today it’s a bare minimum. The best companies grow at 5x, 10x or even more.
Rounds are closed faster than ever — but it’s easier than ever to evaluate the economics of software companies. SaaS diligence is a solved problem.
Funds are deployed faster than ever — but that’s what maximizes dollars returned, not some arbitrary investment schedule.
These arguments are obviously true. But it was also obviously true (and I’m not being sarcastic here) that there were some genuine structural advances in credit markets in the 2000s, broadening access to loans for borrowers while reducing risk for lenders. This did not stop the credit markets from imploding in 2008.
So are these arguments strong enough for venture to be immune to Minsky dynamics? Is it different this time?
Detour: True Risk and Measured Risk
One way to understand Minsky cycles is that they’re driven by the gap between ‘measured risk’ and ‘true risk’.
When you lend money, the ‘true risk’ you take is that the borrower defaults. But you can’t know this directly; instead you measure it by proxy, using credit spreads. Credit spreads reflect default probabilities, but they also reflect investor demand for credit products. A subprime credit trading at a tight spread doesn’t necessarily imply that subprime loans have become less risky (though that could be true); the tight spread may also be driven by demand for subprime loans. Measured risk has deviated from true risk.
Similarly, when you invest in a startup, the ‘true risk’ that you take is that the startup fails. But you can’t know this directly; instead you measure it by proxy, using markups. Markups reflect inverse failure probabilities (the higher and faster the markup, the more successful the company, and hence the less likely it is to fail — at least, so one hopes). But markups also reflect investor demand for startup equity. Once again, measured risk has deviated from true risk.
During Minsky booms, measured risks decline. During Minsky busts, measured risks increase. The flip from boom to bust occurs when the market realizes that true risks haven’t gone away.
The Destination, Not The Journey
So now let’s rephrase the question. Has the true risk of venture investments changed? More rigorously:
Does the compression of timelines in venture change the distribution of terminal outcomes for venture-backed companies?
On that question, the jury is still out. It’s not obvious to me that accelerated markups change the power-law dynamics of venture portfolios. Markups change the journey of a business, but do they change the destination?
If the answer is yes, then there’s no Minsky dynamic at play; what we’re seeing is a rational evolution of the venture industry. Maybe startups are truly less risky now; maybe the market truly has matured. More capital, lower returns, safer investments.
If the answer is no, then venture is very possibly in a Minsky boom, and we’re just waiting for the moment when it turns into a Minsky bust.
What could trigger such a moment?
Reasons For Momentary Lapses
This section is necessarily speculative, but I’ll begin with an observation. There is in fact one well-known death spiral in startup land, and it’s the dreaded down round.
In a down round, a startup running out of cash is forced to raise capital at a lower valuation than its previous financing. This is bad news. Anti-dilution provisions mean that early investors and common shareholders are wiped out. Recent hires whose options are now underwater begin to leave. The startup is perceived as damaged goods, and has to pay above market comp to replace them, attracting mercenaries instead of missionaries. Customers, not knowing if the startup will survive, churn. Finances worsen, predatory investors circle, and further down rounds loom.
A valuation spiral is bad enough. It’s usually accompanied by a talent spiral, which is worse. In a tight labour market, good operators have their choice of where to work. The best startups are able to attract the best talent. Meanwhile, flailing startups tend to fill up with mediocre employees — the ones who can’t find work elsewhere. This makes it even harder to recruit excellent people. The spiral continues.
Down rounds are widely considered the harbinger of doom for venture-backed startups. Understandably, founders and investors go to great lengths to avoid them.
How do they do this? The most common approach is to wait it out. Cut costs, squeeze out short-term revenue, raise bridge loans from less-known investors at the best terms you can, and hope to eventually ‘grow into your valuation’. Sometimes it even works.
This is — not coincidentally — a perfect mirror image of the logic used by many investors today. “I don’t mind paying up; on the current trajectory, even a doubling in price is easily recouped via just a few months of growth.”
But if my hypothesis about time is true, this could be dangerous. If compressed timelines are the driver of Minsky inflows into venture, then anything that delays funding cycles could precipitate a painful reversal. First some startups delay fund-raising because they need to grow into their valuations; then the VCs who invested in those startups have to delay their own fund-raising with LPs because they don’t have the requisite markups; then the LPs reconsider their (hitherto ever-increasing) allocations to venture because the latest returns are uninspiring; and before you know it, there’s an exodus from the asset class. Minsky giveth, and Minsky taketh away.
Comentários