Liquidity Risk and Crypto

Characteristically terrific conversation on Odd Lots this week with Greg Jensen from Bridgewater (here).  He makes the case that the drawdown in US equities so far this year has mostly just been a repricing based on the change in the discount rate.  As a result, the draw down so far is 'overly optimistic.'  

The argument is that we have not yet seen the drop from either deteriorating fundamentals, or rising liquidity risk.  Here's one ominous note, among several:

"You are going into a period where the liquidity hole is getting bigger and bigger....40% of the US equity market can only survive essentially with new buyers entering the market because they're not cashflow generating themselves…that is near a historic high, basically in line with '99-2000... the stocks that need that liquidity are getting hit the hardest, and that is happening quite quickly....you also see that to some extent you constantly need new buyers in the crypto space as well.”    Greg Jensen, co-CIO Bridgewater Associates

Greg has great insights on the market throughout, but his inclusion of crypto here gave me pause.  He says 'to some extent.'  Where is there liquidity risk in crypto today?

First, a step back for the argument in traditional equities market. The equities story of the past four months is relatively straight forward.  Rising inflation led monetary authorities to announce a change in policy, signaling a period of rising rates and an end to asset purchases.  Both changes act to lower the price of equities, which have dropped significantly as a result.

The companies that have the longest time horizon until significant cash flows for shareholders have been hit the hardest, reflecting the greater impact of a changing discount rate.  The longer it will take for a company to earn significant cash for shareholders, the greater the impact on today's value if you start discounting those distant cash flows with a higher number.

Greg’s point is that they the pain of rising rates for companies that are cash flow negative is two-fold. First, in addition to the changing discount rate, there is liquidity risk.  If you don’t have sufficient capital to compensate current shareholders (either with share repurchases or dividends) or finance growth, you are now subject to higher costs for all future capital raises.

All companies that have distant cash flows are subject to the repricing of the first effect; but not all are vulnerable to the second.  A company that has significant revenues but has been investing that revenue into growth that results in low positive or negative cash flow does not have liquidity risk, because it will not need to raise additional financing.  Amazon through the 2010s was an example of such a company – massive revenue growth, but little or no cash flows (or taxable income).

Amazon’s leadership managed the company to optimize for growth, conditional on not needing to rely on outside capital markets for financing.   The operating model is very different than, say, Beyond Meat today, which trades at a similar lofty valuation to Amazon of the 2010s, but which operates with significant negative cash flow, and diminishing cash reserves.  Beyond Meat has seen significant repricing from the change in multiple. Greg is arguing that companies like Beyond Meat will be subject to the double pain of multiple compression and increased financing risk.  He believes that only the former is priced in so far, and such companies will be hurt more going forward than companies that are exposed only to the multiple compression.

The market should of course figure all this out for traditional companies.  But, what about crypto?  Greg includes crypto 'to some extent' as an example of an investment subject to both discounting and liquidity risk.

(Let’s exclude Bitcoin which exists as a potential store of value like gold, not as a potential stream of future cash flow, though it does require constant buyers, at least until ~2140, to purchase newly created coins from miners.)

Most crypto projects are analogous to early-stage tech companies – highly speculative, with a long-time horizon until any significant cash flows, if ever.  They are certainly subject to discount rate risk, and we have seen a significant repricing of assets over the past few months.

Ethereum today generates over $5,000,000 per day in fees.  This is the equivalent of revenue for a company: the customers of the protocol are paying a fee to the protocol for space on the blockchain.  Where does that ~$2B in annualized revenue go?  To further the analogy above, is it sufficient to cover the operating costs, or is more capital needed for growth (e.g. does ETH look more like AMZN in 2010s or BYND today)?

Today, fees paid on Ethereum mostly go to miners. The fees to miners are for security – these fees are to incentivize the miners to record the transaction and keep the blockchain moving. Consider this a cost of goods sold, since it is the amount of revenue that goes directly to the production of the product being bought by customers (block space).

Since August of last year (with EIP 1559), a small portion of every fee paid is also burned. This is the equivalent of a small share buyback from the protocol. The ETH is retired and no longer in circulation. Over 2B ETH has been burned since implementing that change (you can watch these transactions in real time here).

Today, more ETH is still being created by the mining process as the protocol uses proof of work.  This new issuance is equivalent to (very small) initial public offerings of 'equity' every few minutes of every day. Between these issuances and the ETH being burned, there is a net new issuance: about 200,000 in the last month, and over 1.5M ETH since August of last year.

Thus, there is liquidity risk in ETH today.  Every day, there is about 6,500 new ETH ($13M at today’s prices) that needs to find a buyer. To Greg’s point, if there are not sufficient net new buyers each day, the price of ETH will fall.

However, later this summer Ethereum will move to proof of stake for its consensus mechanism. When it does, the fee paid by users will go to validators not miners. Validators of the protocol are ETH holders that agree to lock up their ETH holdings to provide the security mechanism once produced by miners. They are endogenous to the system.  In our analogy, validators are shareholders, so the fee changes from an expense item to the equivalent of a dividend.   In other words, the costs of goods sold for the production of the blockchain product will move to zero, and a new shareholder dividend will be created. (Not all ETH holders will be validators – it does require 32 ETH and a lockup, so a better technical analogy will be a Preferred Shareholder dividend to those ETH holders that choose to be Preferred by becoming validators).

Proof of stake eliminates liquidity risk for ETH.   There will still be a number of other risks, including the risk the protocol will not be able to successfully transition from proof of work to proof of stake.  The market cap of $240B for ETH is still more than 110x the annual fees earned by the protocol.  And, the fees paid to the protocol are for services that are highly speculative (like recording ownership of a NFT).  However, if we follow Greg’s insight on the drivers for equity repricing in traditional markets, ETH should be exposed only to the discount rate change risk and the risk of deteriorating fundamentals, not liquidity risk.  Therefore, a successful transition to proof of stake should meaningfully adjust the overall risk premium for ETH.

In crypto more broadly, the argument above should apply to other L1 proof of stake protocols. Greg's useful framework for disaggregating the types of risk during these turbulent times is a helpful way to consider which assets still have more room to fall.