**We have recently published our first extensive ****research report**** on a digital asset***, which for legal reasons is unavailable to general public. If you are a financial/digital asset markets professional, you can fill out the form on our website with actual (and verifiable) information to request free-of-charge access.*

*In the aforementioned report we present an upgrade to the cryptoasset valuation methodology. Here we would like to share it with a wider audience. We welcome any feedback, especially substantiated critique, as we are firm that the valuation models for digital assets are yet in an early stage of development.*

Cryptocurrency valuation is the key conundrum for traditional investors who have only recently begun paying attention to the new asset class. However, the valuation methods lag, as usual. Equity markets had existed for four centuries and the New York Stock Exchange operated for 130 years before Discounted Cash Flow (DCF) methodology became the mainstream in equity valuation. Unsurprisingly, with 10 years of history no one really knows how to value cryptoassets yet. Despite attempts by a few research enthusiasts, a mainstream valuation method still needs to be developed.

In this article we present our approach which, we believe, is the next meaningful step in cryptoasset valuation methodologies. We assume that the readers are familiar with such relative metric as Network Value to Transactions (NVT) ratio, popularised by Willy Woo and updated by Dmitry Kalichkin, as well as an absolute valuation approach based on the monetary equation of exchange (rooted in the quantitative theory of money), separately proposed by Chris Burniske and Brett Winton.

First, we think that, at least at the current level of public blockchains development, NVT cannot be used to value the networks since the ratio is a function of just one specific variable, velocity. We will derive a simple formula to prove it. Second, we introduce our in-house, based on equation of exchange, valuation methodology which, unlike Burniske’s and Winton’s models, accounts for all the periods of a blockchain development from today to infinity. We compare the results obtained by our approach with the ones received by Burniske’s and Multicoin Capital approaches using INET generic token model.

**Let’s start with the NVT**

NVT is a ratio between a network market cap and its daily transaction volume. Network value (NV) is identical to M (size of the asset base), if we use the equation of exchange designations. In other words, NVT = M / T*daily* by definition, where T stands for aggregate on-chain transactions. Now recall that the right side of the equation of exchange, P*Q, is a blockchain transaction volume by definition also, i.e. M*V = P*Q = T*annual*, with P standing for the price of a good or a service provisioned by the blockchain, and Q representing the quantity of such a resource. Therefore, M / T*annual* = 1/V, where V stands for annual velocity of an asset.

Daily transaction volume, and NVT ratio as a result, are very volatile. To smoothen NVT’s volatility, Dmitry Kalichkin suggested using transaction volume’s moving averages. However, another approach to smoothen NVT would be using a trailing annual transaction volume as denominator in the ratio. We call it NVTannual. In this case, NVT*annual* = M / T*annua*l and we arrive at a very simple formula:

**NVT****annual**** = 1/V**

*Regardless of what kind of transaction volumes we take — daily, annual, moving or any other averages — the inverse relationship between NVT and V will stay: NVT = Const / V. Therefore, a blockchain NVT ratio is a function of just one variable — velocity of its native coins.*

This brings us to some very important conclusions. Given that various coins are likely to have different velocities (e.g. due to different use cases), we cannot use NVT ratios to compare those coins. Moreover, we cannot use NVT ratios to compare the same coin in different stages of its native blockchain development since its velocities are likely to change over time. It seems that NVT is only applicable for mature blockchains with rather stable velocity for its coins. The NVT valuation method would only be used to estimate current intrinsic values of active blockchains, much like P/E ratio can only be used to value profitable entities.

We introduce the concept of the Network Value to Future Transactions (NVFT) ratio, which could be a better approach to cryptocurrency valuations. Traditional finance places more value on price to future earnings ratios than on price to historical earnings. By the same token, it makes sense to use NVFT to value cryptoassets. However, with NVFT even more unknowns appear. The cryptocurrency market has to become mature enough for the analysts that cover blockchains to estimate their future transaction volumes and to arrive at a consensus forecast. So, we will leave it as a concept for now.

Nevertheless, equipped with the NVT / Velocity equation, we can make much more reasonable judgments on NVT ratios. Although without velocity values we still do not know the fair values for the ratios, making assumptions on more tangible velocity is easier than speculating on abstract NVT ratio values. Utility tokens should have high velocity and as such their NVT should be low. If a token changes hands every day, its annual velocity equals 365 and NVT equals one.

**HASH CIB valuation approach**

Now, let’s focus on the main purpose of this article, absolute valuation methods. The fact that Burniske and Winton’s models in reality consider only one, arbitrarily chosen, future period bothers us the most. The model forecasts a blockchain’s CUVs (Current Utility Value) for all the years up to the one when the network matures, i.e. reaches its assumed market share. Then the model discounts the CUV only for one, arbitrarily chosen year, effectively disregarding all intermediary and subsequent CUVs and making irrelevant all the complicated calculations used to derive them. In this case, for instance, a network with initially steep adoption curve would be unfairly valued equally to a network with initially flat adoption curve if both have equal CUV in the year which we decide to choose for discounting.

This model also assumes a constant velocity for the coins, which is unlikely to be the case. However, we don’t consider this a major drawback given that the approach accounts for just one future period and only the velocity in that particular period matters. This also leads us to believe that it makes less sense to model a dynamic velocity if all of the periods except one are ignored, as in Alex Evans’ update to the initial model.

All in all, for a valuation approach to be more robust, we have to account for all the periods of a blockchain’s existence. However, a straightforward summation of all discounted intermediary periods is a pure double counting and fundamentally incorrect, in our view. We have seen this method being used by Multicoin Capital to value 0x and believe that the target price received by their model is unjustifiably boosted.

We introduce our in-house valuation approach which is our modification of the Burniske/Winton model. It considers all stages of a blockchain’s development, and assumes a dynamic velocity, thus, dealing with the above discussed issues. We would hesitate to call the number generated by this approach a “target” or “fair” price, given that the model only applies to coins with utility value and to developing blockchains. We would instead call it a Rational Network Value (RNV).

The rational utility value of a network is not just the discounted future CUV of a particular year, nor is it the sum of discounted CUVs of all projected years, we believe. We think that blockchain’s rational utility value is better modelled as **today’s** utility value **plus** discounted **additional **current utility values (ACUV) for every year to infinity. ACUV*t* in a year **t** equals the difference between CUV*t* in the year **t** and CUV*t-₁* in the year **t-1**.

**ACUV****t ****= CUV****t ****— CUV****t-₁**

To account for infinity, we calculate the terminal value (TV) of ACUV at the period of network maturity. In this case, the model captures all the years. The approach is similar to the one being used to value banks, for which traditional DCF is not applicable. A financial institution is valued as a sum of its current shareholders’ equity and the present value of future excess returns on equity (returns in excess of those required by the cost of equity). A research frontier just a few years ago, this approach has nowadays become the mainstream.

The commonalities and differences of our approach are easier to explain with a simple example. Consider a network that matures in five years and subsequently grows with the annual rate **g** indefinitely. CUV*t* is network’s utility value for the end of year **t**. The additional utility value ACUV*t* for the period **t** is equal to CUV*t *— CUV*t-₁*. TV for the end of the year five equals ACUV₆ /(r-g) = ACUV₅ *(1+g)/(r-g). This is a classic formula for terminal value, where **r** is the discount rate*.*

In this example, the RNV would be as follows:

The charts below show which CUVs are considered in each of the models:

For further comparison, we took Burniske’s generic INET token model and calculated the target prices obtained from each of these three valuation approaches. We didn’t change any assumptions, therefore, Burniske’s approach gives the same original $0.26. Then, we used Multicoin Capital approach and summarized discounted CUVs for all the periods in Burniske’s model to arrive at $3.02. The difference between two approaches gives almost 12x increase in the target price! Using our approach, we received $1.08 per INET token. Still more than a 4x boost to the original target price. However, in this example, we used a constant velocity for all three models. If we used a dynamic velocity in our model, we would have received $0.39 for the target price.

Model with constant Velocity

A cryptoasset’s velocity will most probably increase with growing transaction volume and a blockchain’s approaching maturity. John Pfeffer reasonably argues that a utility blockchain at equilibrium will have a very high velocity and its PQ will equal the cost of computational resources needed to run the network. Even if a fraction of utility tokens is permanently held, for staking in PoS consensus for instance, the rest should change hands often, which would drive up overall velocity.

We use a dynamic and increasing velocity to value cryptoassets in our in-house model. The rationale is simple. Only investors, not end-users, buy tokens during an ICO and on subsequent secondary market, and only investors hold tokens while the blockchain is under development and maturation. Even if they trade, the average holding period is months rather than days, we believe. Thus, velocity is relatively low during the initial stages of the project. If the project is successful and the network is up and running, end-users come aboard, and the tokens are used for their main purposes. As a result, end-users gradually substitute investors as token holders, and velocity grows. In our model we assume that a cryptoasset’s velocity increases concurrently with a blockchain’s adoption according to a logistic S-shaped curve. One can play with the steepness of the S-curve, depending on their assumptions in relation to blockchain’s future success, but the general rule is that velocity and adoption share a similar S-curve.

If we do the same exercise with the three approaches using INET model and the velocity growing from 20 in the year 2018 (as originally assumed in INET model) to 365 (tokens change hands daily) in year 2028, we obtain the following target prices: Burniske’s model — $0,014; Multicoin Capital model — $0.79; HASH CIB model — $0.39.

Model with dynamic Velocity

**Hands-on exercise in HASH CIB valuation methodology**

For final digestion of our valuation approach, we would like to show how it works on an extremely simple model. Since the idea is to clearly explain our methodology, we don’t make complicated assumptions about an imaginary blockchain project, its number of coins in circulation, etc.

*We stick to the KISS principle***.**

Imagine a blockchain project, called UT, which aims to provide decentralized storage service and to take 10% of the global storage market share by 2028.

**Assumptions:**

· Global storage market is estimated to be $30 billion in 2018 and is expected to grow between 22% to 13% from 2018 to 2028.

· UT market share grows from 0% in 2018 to 10% in 2028 according to the classical logistic function (or S-curve)

· UT coins velocity grows from 0 in 2018 to 365 in 2028 concurrently with UT market share according to the same S-curve

· UT has 50 million coins in circulation with a 5% annual inflation indefinitely.

· Discount rate is 40%

First, we derive UT market share for each period according to the logistic function. Second, we calculate UT transaction volume for every year as a product of its market share and the total storage market. This transaction volume is CUV in terms of our valuation model. Third, we calculate the ACUV for every period from 2019 to 2028 and the terminal value (TV). Forth, we derive UT coin velocity for each year. Then, we find the number of coins in circulation for each period. Six, we compute the discounted ACUV for each period and TV for 2028 and divide each of them by the corresponding number of coins as well as by velocity values. Finally, we sum up CUV for the initial 2018 year and all the values derived in the previous step to arrive at the Rational Value for UT token at the end-2018.

Generic HASH CIB model (KISS ver.)