Crypto assets is an entirely new asset class that was created over the last few years. As more people try to get into the cryptocurrency market, there is more than ever, a need for a reliable methodology to accurately calculate the value of a crypto asset. The efforts to attempts to qualify and value crypto assets is indeed laudable, but they suffer from a few flaws…

Firstly, there is a severe lack of empirical analysis… Any prediction model will require rigorous empirical evidence, and the crypto asset market is still fledgling, making it extremely difficult to test these models due to the lack of necessary data. This is further compounded by the diversity of the crypto assets in the market, for example, asset-backed securities tokens, utility tokens, and cryptocurrencies, etc.

And this leads us to the second problem, the lack of consistency in the type of data being used… How are the fields defined? And where is the data coming from? Crypto tokens have currency-like properties such as the medium of exchange and store of value, along with other functional objectives. Therefore a clear definition of the type of data to be applied is critical to the development of rational valuation models.

Thirdly, there is a huge number of assumptions that are being made in the development of such valuation models. And the single biggest issue is that such models are based on traditional stock valuation are now used to value a new asset class that has different inherent properties from stocks and shares. Very few token valuation models built are able to explain the assumptions behind the formulas before applying them to the new instances.

And finally, there are concerns about overfitting and model complexity… Data scientists are concerned with Bias-Variance Tradeoffs, which states that the higher the complexity the more variance it causes and leading to overfitting. Finding an optimal balance is really challenging. And this issue comes back to the first problem of the lack of empirical evidence and can thus cause invalid correlations.

To truly understand crypto asset valuation methods, we need to delve deeper to the history of valuation models and methods. Early financial markets went hand in hand with fraud. Since the creation of the first stocks and trading market, the ability to distinguish between the intrinsic value of an asset from the speculative value of the same asset is of utmost importance to any investor.

One of the first people to address this need was Benjamin Graham in 1934. Graham offered the first formal approach to the determination of the intrinsic value of a stock. And it is this work that leads to the birth of financial analysis and corporate finance. Graham stated that intrinsic value is that value which is justified by the facts, e.g. the assets, earnings, dividends, definite prospects, etc, from the market quotations established by artificial manipulation or distorted by psychological excesses. By driving a wedge between intrinsic value and speculative value, Graham’s work allowed investors to determine what the underlying value of a security is, how to determine it and how to apply that in speculative instances.

With a token valuation, it is this exact core principle of financial analysis which is missing. Several efforts have been expended in trying to find a logical model to value crypto assets. But much of it are attempts to squeeze a square peg into a round hole, recycling stock valuation models to create a new valuation method for crypto assets.

Crypto tokens are not the same as stocks… A crypto token has similar properties to a currency, whereby it functions as a medium of exchange and as a unit of account. So, therefore, it shares at least two, if not all three properties of a currency: the store of value, a unit of account and a means of transfer. And this means that a crypto token has to be evaluated both as a securities and currency at the same time.

A crypto token is also used to create, generate and stimulate value. It functions within an interconnected network such as a blockchain and is also an endogenous monetary system with its own demand, supply and liquidity issues. To evaluate such systems, one needs to look at the monetary policy as well as the underlying business behind this network.

Most importantly, there is a big question of diversity… Today, there are many different kinds of tokens being issued, and each category has its own distinct characteristics and properties. Some of the common ones are securities tokens, utility tokens, payment tokens and so on. This diversity, coupled with the supply and liquidity issues further complicate the matter. Appended below is a chart that shows the current taxonomy of crypto tokens and helps us visualise the diversity of this space.

So let’s now look at some of the more popular valuation models for crypto assets now…

Token Velocity Methodology

Token Velocity methodology is one of the most popular valuation models in use now due to its nuance in considering a token-based economy as a monetary system. In doing so, the straight forward application for its application are independent cryptocurrencies like Bitcoin, Dash, Monero, etc. The main idea behind this valuation method is that token transaction velocity is one of the key levers that determine long-term token value.

Drawing from The Monetary Equation of Exchange (MV=PQ), which is often referred to as the Quantity Theory of Money, velocity is a significant driver of token price. The lower the velocity, the greater the token price is via an appreciation of M on the left side of the identity. The implication of this thesis is that tokens with low velocity, i.e. those that sit longer in wallets for whatever reason (owing to speculation, asset-backed, etc.), will see higher prices than other coins, all else equal.

The valuation method has some fairly widely recognised criticisms:

M itself is very difficult to measure in the crypto space, as there can be locked up or un-mined currency that may or may not is reflected in the model’s M value, making it extremely difficult to establish the size of the asset base.

V is assumed to be a constant, which is hard to calibrate within a functioning economy whose natural state is entropic rather than equilibrium. The nature of the token economy thereby makes it rather hard to precisely define or measure Velocity.

The other factors in the equation, M, P, and Q can also not be easily measured or estimated. In fact, economists would say that you need models to estimate any one of these variables along with their correlations with one another.

When velocity changes, the choice to record the effect in M, P or Q is arbitrary and yields different implications for a token price. Further, V’s relationship and correlation with these factors are dynamic, and assuming a steady relationship with P, Q, or M is again arbitrary and problematic.

Crypto J-Curve Methodology

This method is a recent idea and is an extension of the MV=PQ approach. In this model, a token’s price is based on two components whose contributions to the token price evolve over time. The CUV refers to the “current utility value”, which represents value-driven by utility and usage today, and the DEUV refers to the “discounted expected utility value”, which represents value-driven by investment speculation.

A token’s current market value can be modelled and projected using inputs including supply-side drivers, adoption, and market saturation growth rates, token demand, and velocity. Further, CUV and DEUV and their respective dynamic influences on token price can be modelled and estimated. Following the monetary equation of exchange (MV=PQ), the token price equals the projected monetary base (M) in the future divided by the number of coins in circulation in the future; M is calculated as equal to PQ/V, or the value of on-chain transaction volume (or “network GDP”) divided by token velocity.

Over the token project’s development cycle, the CUV and DEUV take turns driving token prices as the projects and the market perceptions of the tokens stabilise and mature. When a token is first launched, DEUV dominates as holders are excited about the technology and expect appreciation of future price. When enthusiasm wanes with inevitable technical roadblocks, the price declines and is driven more by CUV from technical users and early adopters. As the team overcomes challenges, CUV quietly grows as the token becomes more widely adopted. DEUV then catches up as speculation and excitement follow developer interest. Ultimately in the steady-state, CUV should drive token price.

Some adopters of this Crypto J-Curve method have begun to use this as a measurement of the different stages in the life cycle of a crypto asset, from the ideation to the release to the development to the deployment of the project.

Network Value-to-Transaction Ratio (NVT) Methodology

This is an interesting method that was developed quite recently and is an adapted version of the stock valuation method Price-to-Earnings Ratio (P/E Ratio). This valuation ratio compares the network’s value (the market cap) to the network’s daily on-chain transaction volume.

NVT = Network Value / Daily Transaction Volume

NVT may indicate whether a network token is under or overvalued by showing the market cap relative to the network’s transaction volume, which represents the utility that users derive from the network. When the ratio becomes very high, it indicates potential token over-valuation. The model is interesting as it the first that looks at network attributes rather than financial models. The ratio best applies to assets whose on-chain transaction volume closely represents utility to users.

Here are some of the common criticisms of this NVT Model… Transaction volumes tend to follow changes in price. The higher the price, the greater the tendency to store the token and not use them. Thus these two variables have an endogenous and “reflexive” relationship, weakening the indicative power of the ratio. There have been some industry experts who are experimenting with the time frame used to measure daily transaction volume.

Apart from the methods listed above, there have been other approaches that have tried to explore adapted models that use metrics such as EV/EBITDA, P/E, EV/Sales, Carhart four-factor CAPM model (Crypto CAPM), Sharpe’s ration and Black-Scholes Options Theory.

Crypto markets are very new with limited data history pertaining to crypto asset behaviour, returns, and correlations. Many of today’s models are simplistic or limited, whether intrinsically (due to difficulty defining and measuring variables such as velocity and its counterparts, for instance) or extrinsically (due to limited applicability to different types of tokens, as seen with NVT and privacy coins, for instance).

In the future when the markets mature and asset relationships and behaviours are more discoverable, valuation models and ratios should be more predictive and informative. However, because of the very diverse nature of crypto assets, which can have different features, structures, payouts, etc., we may never have metrics and models as universal as the P/E ratio and DCF analysis for public equities.