NFTs, Liquidity, and the Attention Economy (Part 1)

Hey everyone, I’ve briefly returned (or more accurately, I’ll write posts when I have a free minute). I’ve kept joking this past year that the market could not get any stupider, and then it always seems to:

Many people have some weird canard they like to ascribe to this regime shift, picking whatever the Boogeyman of the Day is - quantitative easing, stimulus checks, increased unemployment, you name it. I’m not really here to debate the macroeconomics of it all or figure out a way to trash social welfare programs, but I do think that the end result is fascinating. As I’ve discussed before, valuation can be understood as an expression of belief, and codified not in dollars or some arcane fundamental view, but in regret. This helps make sense of the current market and is fairly resistant to macroeconomic trends—if the marginal value across all market participants of a dollar decreases (e.g. increased unemployment benefits), we’d expect that the dollar value may increase while regret (risk in real economic assets) remains the same.

Before someone accuses me of Dunning-Kruger for the previous sentence, the issue lies in understanding stylized facts — key empirical findings that essentially will make or break a model’s validity. If we subscribe to a fundamental view of valuation and castigate those straying from the light of DCF as “irrational”, we miss an entire universe of rich market findings. If the market strays so far from a fundamental view of value, is it the market or our models that are wrong?

The most extreme slippery slope of this argument is that perhaps value as a concept is unknowable, similar to the old legal axiom of obscenity in that “I know it when I see it”. I affectionately refer to this as the epistemological nihilism school of valuation—that valuation is not a quantifiable (or perhaps even qualifiable) concept, but unique per case.

I do not really believe that valuation is some unknowable truth, though, at least as a heuristic. I don’t think any model of valuation will be complete for all assets, though. My own perceptions of value have changed over time, and I’m sure they will continue to change.

For many of us recently, our idea of valuation has been shaken up a bit more (even in comparison to meme stocks and the whole concept of cryptocurrency) by the advent of NFTs (non-fungible tokens), which again exploded in popularity this summer. Many takes have been had on them:

For those not in the know, NFTs - non-fungible tokens - are one of the natural uses of the cryptocurrency blockchain — they essentially store information pointing to a content resource (or in a few projects, store the actual content). This is very much akin to a web domain address pointing to a website, with the added benefit (or cost, depending on your perspective) of being non-fungible. What does that mean? Let’s take this website, When a user types in that URL into their browser, they’re taken here:

That said, this relationship is fungible — if Substack decides to ban my blog it’s possible the URL will eventually instead point here:

This means philosophically that I cannot equate ownership of the blog “Midnight on the Market Momentum” to simply knowing I am the owner of - those are related but very separate things. An NFT, being non-fungible, is different. In the most common permutation, an NFT represents a location to retrieve content like our Pudgy Penguins:

There are two things critical to note here:

1) The ownership of an NFT is represented by a locator - in most NFTs, the content represented by it is stored on a server somewhere, where the NFT represents a link, but not the actual content. So if the server holding it goes offline - too bad.

2) An NFT is represented on one blockchain - this is moderately key to note. If one owns the rare fellow represented above on the Ethereum blockchain, there is no “legal” reason why an industrious grifter might not mint an identical content NFT on the Solana blockchain. This is a known issue, and many are working on the solution.

However, this article isn’t really meant to talk about the technical specifications of the NFT ecosystem. In all truth, I do not really see a strong value proposition for the existing NFT ecosystem—but that may (or may not) be on me. It’s clearly a speculative bubble, but I’d be hard pressed to know what the victors and losers of the bubble will be— it’s always much easier in retrospect. Regardless of your opinion of the utility of NFTs, in recent months, we’ve seen eye-popping valuations on some of the better known projects:

Given an exchange rate of $3,272 per Ether (at time of writing), this works out to nearly $58,000,000 worth of trades to purchase and sell Pudgy Penguins, and a “floor price” (the minimum for-sale price for an NFT collection—a bit disingenuous wording) of $7,852 minimum to buy a link to a JPEG of a cartoon penguin.

It’s very difficult with a straight face to talk about valuation here in a fundamental perspective because, to wit, there is none. This is a Known Problem in art valuation in general, for a few reasons:

1) Liquidity - In general, the most simple metric of value is current market price, which in even simpler terms is how much someone else will pay for it if you sell it right now. In a deeper market with commoditized units (e.g. there is no material difference between shares of the same class/company, or sheets of metal according to the same specs) it tends to be much simpler to derive current price, since transactions are largely continuous. However, in largely illiquid markets (housing, art, wine, etc) two factors emerge — in general the assets are not commoditized (one wine vintage is not the same as another), and prices are not at all continuous (for the same artwork or even the same collection, transactions may be years apart). In more illiquid markets, there is significantly higher sensitive to macro trends, versus microstructural phenomena (which, in the absence of news, may largely dominate intraday pricing of liquid markets).

2) The nature of art value - Art is, to some degree, both objective and subjective. The subjective aspect comes from the personal “fundamental” value of art — at a basal level, there is an intangible value we can ascribe to acquiring art we deem aesthetic or meaningful, regardless of financial consequence. This is an experiential premium, and is purely rational, in the same sense we invest money to go to Disneyland or on a nice vacation. Unlike economic value, though, this experiential premium largely differs on a personal basis, and it’s extremely hard (if not impossible) to measure. Each individual participant has some likelihood (not certainty) of their personal valuation and the quickest measure tends to be “maximum price one is willing to pay”. If I wanted to know how much I “valued” a trip to Disneyland, I could simply figure out the maximum price I’d be willing to spend to go there ($100 a day? $200? $1000?). As a rational actor, given I cannot resell this experience, therefore all value (and hence my willingness to pay) should be a function of the utility I derive from it.

What’s interesting as we start thinking of a generalized framework for understanding asset valuation is we can largely separate valuation into two components:

1) Fundamental value — Fundamental value isn’t necessarily discounted cash flows. Fundamental value, in this perspective, could be seen as the innate value derived from an asset removed from its liquidity value. In a salient example, a toothbrush has a “fundamental” value in the sense that you need it to brush your teeth to maintain dental hygiene, regardless of if anyone else in the world has any value for it at all. A vacation to Disneyland has value even though it isn’t transferrable (after consumed, of course) because of the beautiful memories you create and the endorphins you feel. From a fundamental perspective, we can see multi-dimensionality largely depending on the asset class (this is not an exhaustive list):

  • Experiential/personal value

  • Economic/industrial value

  • Net present value

For traditional asset classes like equities and bonds, we expect that a rational actor should be indifferent to the underlying issuer itself, and only care about the net present value, which largely is a function of discounted cash flows. This tends to be much easier in Asian equities (e.g. the Hang Seng, where issuers are represented by numbers, not names), but clearly breaks down in the age of narrative investing.

In this perspective a cult asset like AMC or GME or even Bitcoin may have a fundamental value that may be substantially larger than its market-implied NPV, given the personal value premium many shareholders attribute to it. The interesting impact of that conclusion (the added premium of narrative/personal value) is that while the market price is fairly liquid and susceptible to change, the personal value (largely being driven by group psychological phenomenon) is much more “sticky” and persist over longer periods of time.

This carryover from illiquid finance seems wrong at first, since it breaks the idea of rationality. That said, the beautiful thing about reality is it doesn’t exactly obey the neat mathematical models we want to believe it does. And when it doesn’t, is it reality that’s wrong?

2) Liquidity value — Liquidity rules everything around me, or so they say (especially for the group of folks who use liquidity as the boogeyman without knowing what it actually is). It’s a well supported theory of economics that we condition assets with a liquidity premium. Let’s define what liquidity means here. Liquidity here is the ability to sell or buy the asset at any time without incurring much market impact - adverse price movement. In a market of willing buyers and sellers—ergo, if I own an asset and have no rush to sell or buy—the asset price should arrive at the fair market price.

However, unless the price of the asset is precisely its fundamental value (e.g. when buy a toothbrush to use it ourselves), we have to decouple liquidity value into two factors:

1) A liquidity premium for forcible liquidation — The most discussed variant of the liquidity premium theory is forcible liquidation, or the demand for cash.

This is well known as the liquidation value of an asset for any given time period - if I needed the money for my house tomorrow, how much can I get for it? In general, the less demand there is (one of the two components of an asset’s liquidity, along with supply), the more adverse price will be on immediate liquidation. In the most degenerate case, if I need the money for my house tomorrow and there is exactly one buyer (and I cannot take a loan on it or some form of partial liquidity), I will have to accept any offer they give me. Therefore rationally, given the choice between an illiquid and liquid asset with the same properties (our fundamental value described above), I should pick the more liquid asset.

2) A liquidity premium for uncertainty of future market composition — Even in a scenario where we may have no immediate demand for selling or buying the asset, there exists a non-zero probability that there will never be a future buyer or seller available. This is of course an extreme case, but is much more impactful in decision of illiquid, hard-to-objectively value assets (e.g. art), where the fundamental value is less an objective than a subjective truth. In traditional equities or bonds, regardless of the existence of future sellers (if I am holding the equity or bond), I can rely on the fundamental value as an objective gauge of “value” — I will receive dividends or coupons accordingly. Or to phrase it more succinctly — assuming no need for future, immediate cash, I should be indifferent to the expected value of all its future cash flows and the amount I paid for it (if they are equivalent).

However, the above only works in very limited asset classes that provide an objective measure of return. In many instances (like in growth equities), there may be zero or heavily uncertain future cash flows from the asset, yet it undeniably has a market value. If you were to offer me a Picasso for free, I would rationally take it, even though the Picasso will never produce an income stream for me. Even if I personally valued the painting at $0 (as in, I have no soul), the expectancy for me is positive (especially if there are zero carrying costs), given I can expect a future buyer to value the painting above $0.

That said, the existence of this “future buyer” is not certain. In the painting example, there is a clear carrying cost (storage and preservation), so if I personally do not get value from the Picasso, I would rationally only accept the gift if the expected value of a future sale outweighs the total carrying costs I pay until the point of sale.

Here, even if I have no immediate demand for liquidity, I do ascribe a value to any liquidity at all, from now to the infinite future (or more realistically until I die). I would only derive value from holding the Picasso based on the premise that at some point (whether tomorrow or at the heat death of the universe) there exists a buyer who is willing to pay me. More over, the amount this buyer must pay me for the deal to be “worth it” is proportional to the time I have to hold the asset (given the carrying costs of holding the physical asset). If this buyer never materializes, I should not accept the gift even at zero cost — it becomes my white elephant.

We can loosen the nomenclature a bit, continue this train of thought, and perhaps give it a bit of mathematical notation. When we buy an asset, we aren’t actually giving a premium to all liquidity — we’re giving a premium proportional to our risk of accepting a liquidation value below our paid value (the regret principle). In the extreme case where our value of the asset solely comes from the expectancy of future sales, in the case of no future buyers we should value the asset at zero (or negative if there are costs attached). The future is uncertain, though, and except in rare cases we can never objectively say “this asset will not have any future buyers” unless it is fully consumed (we’re not at the point where you can resell vacation memories, yet).

This has some interesting ramifications.

Even if we ascribe a fundamental value of zero to an asset (based on whatever flavor you believe in), we can define the liquidity premium as the expected profit of selling the asset in the future at any given point (subtracting out the amount you paid and carrying costs). For a speculative asset (often called a bubble), there exists a high degree of uncertainty in future market composition, and a strong degree of inference that needs to be made on fairly limited data. What’s interesting here is that market irrationality can arise from the coordination of fairly rational actors — even if the fundamental value is zero, based on observation we’d anticipate a liquidity premium to occur during the hype cycle. This isn’t a new concept—it is often referred to as “Greater Fool Theory”. I tend to dislike this term, because it has derogatory implications — given the existence of the liquidity premium, we can preserve rationality even if the aggregate outcome might be irrational.

Well, that was a long post. In the next post I’m going to get a bit more into the weeds about some toy mathematical models, liquidity, speculation, and the interesting ramifications of NFTs and other digital assets.

Cheers and stay frosty.

  • Lily Francus