22. The Cryptoeconomics of Token Design
While a cryptocurrency is originally invented as part of the blockchain consensus mechanism, subsequent developments have proven that the cryptocurrency is critical for driving blockchains’ adoption. A blockchain ecosystem’s network effect depends strongly on the cryptocurrency design, which incentivizes how contributors and consumers interact with each other on the network.
In this chapter, I will discuss cryptocurrency (token) designs, known as cryptoeconomics. By understanding cryptoeconomics, you will gain a better understanding of exactly what types of applications are suited for blockchains.
There are three broad categories of cryptocurrencies (tokens): network utility tokens, application utility tokens, and security tokens. Some cryptocurrencies can simultaneously belong to multiple categories.
Network Utility Tokens
The blockchain establishes collaboration among trustless peers. It can offer “trust as a service,” ranging from a secure ledger to enforcement of smart contracts (i.e., guaranteed execution of certain software code) to transparent record keeping. Network utility tokens are used by blockchain users to “pay” for such network services. Users acquire and spend the tokens because they get value and utility from the previously mentioned trust as a service.
A blockchain is a decentralized network, and there is no corporation in the middle to issue orders and give out salaries. The rules and protocols of the blockchain network must be maintained and enforced by the community members (i.e., the contributors) in exchange for tokens. Contributors run computer hardware and software to support blockchain nodes, and they participate in the consensus and governance processes. These contributors are known as miners (in proof-of-work [PoW] consensus blockchains) or validators (in proof-of-stake [PoS] consensus blockchains). You can read more detailed explanations of PoW and PoS in Chapter 2.
The miners and validators receive tokens for creating new blocks as part of the consensus protocol. They are also “paid” to perform computation to validate the transactions in a block or to execute smart contracts involved in the transactions. This transaction fee is typically paid by parties originating the transactions. As a blockchain matures, the miners and validators should mostly be compensated by fees. This forms a closed loop where people who provide services to maintain the network (miners and validators) receive fees in tokens, they sell those tokens in exchanges, and people who consume network services (users) buy tokens from the exchanges to pay the fees (Figure 22.1).
Figure 22-1 The closed loop of a blockchain economic system
Initially, the network utility token has no or little value. As the blockchain network itself becomes useful, more and more people want to use the services the network provides. The fundamental value of a crypto token could therefore be tied to the value of the utility provided by its underlying blockchain network (see the fat protocol theory discussed in Chapter 1). If lots of people are willing to pay to use the network, the crypto tokens will in turn have significant value. Next, let’s look at some examples.
Bitcoin (BTC)
Narrowly defined, the Bitcoin network’s utility service is to securely and transparently record digital transactions. Broadly defined, however, the utility is to provide a trusted store of value.
Prior to Bitcoin, all publically available digital tokens are infinitely duplicable and hence are useless as a store of value. Today, Bitcoin becomes a store of value (i.e., the Internet gold) because it can be cryptographically stored and moved around. So far, no one has been able to hack the Bitcoin system by recording fraudulent transactions. So, trust is a chief utility provided by the Bitcoin blockchain. It is the same kind of utility that gold or diamond provides for society. Likewise, the valuation of BTC is often compared to the world’s gold reserve.
As people become interested in using BTC as a secure store of value (i.e., the Internet gold), they pay for the network’s bookkeeping services in BTC via transaction fees.
Ethereum (ETH)
The Ethereum network’s utility service is to provide a trusted platform that guarantees execution of computer code (known as smart contracts). When parties enter into a smart contract on the Ethereum blockchain, both can be assured that the contract will execute as the code is written. This level of integrity is provided by Ethereum community contributors (miners and validators) who run Ethereum nodes. Those community members are paid by the Ethereum crypto token ETH. As more people are interested in using the Ethereum blockchain to enforce their own smart contracts, the demand and value of the ETH go up.
Interestingly, like the BTC, ETH is also increasingly viewed as a store of value but for a different purpose. As decentralized finance (DeFi) applications, such as the Uniswap Exchange, MakerDAO stable coin (SC), and even initial coin offering (ICO) fundraising, take off on the Ethereum network, the ETH is increasingly used as collateral for financial investment. ETH holders are receiving interests, dividends, or “rewards” for their ETHs locked in collateral pools. The ETH has become an investible asset that can generate returns. As we will see later in this chapter, this reduces the velocity of ETH circulation, creating a virtuous cycle for ETH valuation.
There are many efforts to improve the Ethereum blockchain, including the EOS, QTUM, ATOM from Cosmos, and CMT from CyberMiles. Their crypto tokens have similar utility values as the ETH.
ZCash (ZEC)
The ZCash network’s utility service is to record anonymous and encrypted transactions so that no one can figure out the parties involved in those transactions. Privacy-concerned users use this utility service to conduct transactions and must pay the network’s community of maintainers fees in ZCash for each transaction.
Application Utility Tokens
As discussed in Chapter 1, blockchain networks have the potential to replace corporations as a way to organize production of digital goods and services. Besides trust as a service, a much bigger economic opportunity is for community members to contribute application services to be sold on the blockchain networks. The blockchain aggregates service offerings and conducts buy/sell transactions in crypto tokens. The value of the token can be directly mapped to the value of the aggregated services provided via the blockchain. Let’s look at the example of a hypothetical storage sharing coin (SSC) to illustrate how an application utility token might work.
The SSC blockchain provides cloud-based data storage space for users. The storage space is provided by millions of community members who contribute spare hard drive space and Internet bandwidth from their own computers. The blockchain aggregates the fragmented storage space and offers to sell them on-demand on the Internet.
As users demand and use the cloud-based storage provided by the blockchain, they must pay with the SSC token. The blockchain accepts SSC tokens, provides storage space, and distributes SSC tokens to storage space providers. All these are done through protocols and rules codified in software and enforced by blockchain consensus.
The SSC does not necessarily require its own blockchain. It could be implemented as a series of smart contracts on the Ethereum blockchain. The smart contracts specify economic parameters of the system including token supply, fee structure, use of tokens, distribution rules, and so on.
Obviously, the SSC represents just a tip of the iceberg. Besides data storage, community members can provide a variety of useful digital products and services, including intellectual properties, personal data, solar-generated electricity, and medical records.
The common attribute here is that those digital products are valuable in the aggregate. The essential function of the blockchain is to aggregate them from trustless members of the community and fairly distribute the profits. In turn, users are required to pay for the aggregated product and services in the blockchain’s native crypto tokens.
Security Tokens
One of most exciting uses of crypto tokens is to represent traditional ownership securities, such as stock shares or even shares in a house or a car. Unlike today’s securities, this new form of security is programmable and enforced by smart contracts on blockchains. It opens many compelling use cases in the world of artificial intelligence.
For example, let’s consider a self-driving car that acts like a taxi or an Uber car. It makes money and delivers its profits to its owners or stakeholders. With programmable tokens, you can receive the car’s profits when certain dynamic conditions are met: when the car is up-to-date on insurance, is traveling in areas and speeds you designate, and is only picking up passengers of a certain profile. In turn, your token will also be responsible for liabilities and losses incurred at those times. A different stakeholder of the same car might want to profit from a riskier business strategy and get paid when the car is taking on more risk. Here are some examples:
When the car is driving during the day and when the traffic is normal (low risk of collision), owners of token A will receive the profits.
When the car is driving at rush hour for higher fares and subject to a higher risk of collision, owners of token B will collect the profits and be responsible for the increased risk of loss.
When the car is driving at night in a high-risk neighborhood to pick up drunken passengers, it will receive the highest fare and has the highest risk of damage. Owners of token C will collect the profits and be responsible for the potential damages.
Today’s security regulations are designed for the era of “dumb” shares and bonds. The regulations still need significant updates before the vision of smart programmable securities can be realized. However, there have been many innovative attempts. Here are some examples.
The DAO
The decentralized autonomous organization (DAO) experiment by Ethereum was to raise an investment fund governed by smart contracts in terms of investment decisions and profit distributions. While the effort was unsuccessful because of technical issues, the idea was illuminating. The Securities and Exchanges Commission (SEC) of the United States reviewed the Ethereum DAO and decided it was issuing securities—albeit a programmable, smarter, and much more transparent security than regular shares in a fund.
Token Funds
Multiple traditional venture capital (VC) firms have raised new funds through the ICO mechanism. Examples include the Blockchain Capital fund (http://blockchain.capital/) and the Science Incubator (https://www.science-inc.com/). The stakeholders in those funds are no longer known as limited partners (LPs) but as token holders who have instant liquidity in their shares and will get investment returns through the crypto tokens they hold.
Because of the huge market potential in this nascent market, we recommend you closely follow the developments in this space.
Token Valuation
The use of crypto tokens to consume blockchain services gives valuation to those tokens. In reality, the token price is a major factor in driving adoption and building network effect for a blockchain.
At the time of this writing, a common approach to valuing blockchain networks is to simply multiply the unit token price by the total number of circulating tokens to get a market cap (see http://coinmarketcap.com/). This is analogous to valuing a company’s market cap by its stock price. However, as we noted in Chapter 1, a blockchain network is very different from a company. For one thing, the blockchain network is not-for-profit and the price-to-earning (P/E) ratio is meaningless. There is no conventional “sales” measure—only transaction volumes on the network (analogous to the gross merchandise value [GMV]).
We believe that a more appropriate approach to value blockchain networks is through their economic output, similar to how national currencies are measured by the economic output (i.e., gross domestic product [GDP]) from nations. On a blockchain network, there are service providers (i.e., miners, validators, application service providers) and consumers. The crypto tokens are designed to facilitate transactions between the parties and, more importantly, to incentivize collaborative interactions in the ecosystem.
Utility Tokens
For network and application utility tokens, macroeconomics theory indicates that the price of each token can be determined by the following factors. The following formula is known as the value of exchange equation:
P is the price level, which is the price of services in terms of the token. So, the token price is actually 1/P.
T is the total value of services or products the blockchain network community produces in a unit of time (say one year). The long-term value of the token is determined by the value of the underlying services the blockchain provides. In that sense, the token is “backed” by the value of the services.
M is the total supply of tokens available to exchange for such services and products.
V is the velocity of money (monetary velocity) as measured by the number of times an average token changes hands in the unit of time. It is inverse to the average time token holders hold a token. The higher the speed, the lower the token valuation since each token can be reused to purchase the set number of products and services.
Compared with money supply and velocity in the traditional monetary system, we can estimate V based on how the money supply, M, is utilized in the system.
M0 refers to cash money in circulation.
M1 refers to highly liquid and available money, including M0 and checking accounts and traveler’s checks. In the United States, M1 USD has a velocity of 5.6 for 2019.
M2 refers to less liquid money such as savings accounts and money markets. M2 money often carries the purpose of store of value as well as medium of exchange. The USD M2 velocity is around 1.5 in 2019.
As savings accounts dramatically decrease the M2 velocity of USD, the velocity of money is a parameter that can be designed into the blockchain protocol. Here are some examples:
BTC is often used as a store of value, and hence users tend to hold it for long periods of time. That gives BTC an extremely slow monetary velocity and therefore is very valuable. In other words, people tend to hold BTC for long periods of time, which severely limits the BTC supply on the market, causing its price to go up.
A utility token for smart contracts (e.g., ETH and CMT) has a natural holding time because of the design of smart contracts. For example, in escrow contracts, the tokens must be held in the contract account for days before the escrowed transaction can settle.
In an e-commerce blockchain application, buyers often purchase tokens in bulk, and the sellers wait until they accumulate significant amounts of tokens before cashing out. They do that to minimize exchange fees when converting from crypto tokens to fiat currencies (e.g., U.S. dollars).
In an application utility token like the SSC, the blockchain network could require service consumers to deposit and keep a balance of several days’ worth of tokens to ensure uninterrupted services.
Notice that the price computed via the exchange equation represents the current intrinsic value of the token. The actual trade price of the token should reflect people’s expectation of the token price in a few years. However, as future money is less valuable than today’s money, we will need to apply a discount. Let’s assume that the market size is growing at gr percent per year, considering both overall market growth and the blockchain ecosystem’s growing penetration of the market; assume the discount rate is dr percent per year. The discount rate, dr percent, is typically between 5 percent and 10 percent depending on the risk of the market. This is the interest rate you incur for borrowing money in this market. After n years, the token price can be discounted to a present value as follows:
T0 is the blockchain network’s GDP in the current year (year 0). Mn is the number of floating tokens in year n, which depends on the economic system design. Typically, we use n = 5 for computing the present value of the token.
Design Considerations
In general, the more value-add the blockchain network provides, the better it can influence users to hold tokens and decrease the velocity of money. That is why we consider blockchain networks that can aggregate smaller service providers more valuable than ones that simply match buyers and sellers.
A common and rookie design flaw is to build appcoins that act purely as a medium of exchange. That is, the service providers and consumers use the tokens only during a transaction, exchanging the token from and to other currencies immediately before or after the transaction. Since no one holds the token, there is always a glut of such tokens for sale on the market, driving its price down to zero.
Nobel Laureate in Economics Paul Krugman once said, “To be successful, money must be both a medium of exchange and a reasonably stable store of value.” As we discussed earlier, an appcoin that is simply used as a medium of exchange is unstable and could crash to zero. The store of value requirement is to give users a reason to hold the currency in between transactions and hence maintain a stable monetary velocity for the system.
However, it is also a chicken-and-egg problem: the currency can become a store of value only if it has a stable monetary velocity. In the case of Bitcoin, the early miners decided to hold it hoping the price would rise in the future. That reduced the monetary velocity and helped to establish Bitcoin’s “Internet gold” status as a value store. For most utility tokens, a well-designed protocol could make holding part of the utility (e.g., staking and voting, or a reserve system) and create a value store.
An Alternative Approach
An alternative approach to apply the value of exchange equation is as follows. The benefit of this alternative approach is that it figures out the monetary base of the crypto asset in USD. If the token serves more than one purpose (e.g., both a security token and an application utility token), we can derive the asset base in two ways and then add them together, which we will show here:
M is the monetary base of the crypto asset. In other words, it is the total USD value of the tokens. It is the “market cap” of the tokens.
P is now the price level of the service provided on the network. This price is USD per unit of service.
Q is the quantity of the service available on the network. Hence, the product of P and Q is the total GDP of the network ecosystem.
V is the velocity of money in the same period when the GDP is computed.
Similar to the method we used before, we need to discount the future monetary base of the assets back to today’s value to take into account the market’s future expectation. Let’s assume that the blockchain ecosystem is growing at an annual rate of gr percent and the discount rate is dr percent per year (with dr percent being between 5 percent and 10 percent depending on the risk). The present value of the monetary base is as follows. P and Q are both present values.
To compute the price for each token, you can divide M by the number of free-floating tokens at year n. Again, we typically use n = 5 for a reasonable future forecast.
Now, imagine that a token can be used both as a security and as a medium of exchange. The token’s market cap value as a dividend-earning security can be computed using the traditional security asset evaluation methods, such as using the discounted cash flow (DCF) methods by discounting and adding all the free cash flow generated from the ecosystem in the future. On the other hand, the token’s market cap as a medium of exchange token can be computed from the previous equation. Those two market caps are additive.
We can then divide all tokens in circulation into two categories: earning and payment. If a token is used 30 percent of the time as a payment token and 70 percent of the time as a dividend-earning token, we will put 0.3 token into the payment pool and 0.7 token into the earning pool. The overall token price reaches equilibrium when the price of each use cases matches each other.
EP is the equilibrium price of each token.
Ms and Me are the present values of the tokens from dividend-earning security and exchange uses, respectively.
Ns and Ne are the numbers of tokens primarily used for security or exchange purposes, respectively.
It is easy to demonstrate that when the token is only a utility token, this alternative method gives the same token price as the formula we used for utility tokens. We simply assigned alternative meanings to components in the value of exchange equation. This new method is easier to apply in situations where there are multiple uses of a token.
Advanced Topics
Now I have covered the basics of token economics. In this section, I will discuss some more complex topics in token design. I will not get into the technical details but aim to provide a glimpse into the world of cryptoeconomics research so that interested readers can explore further on their own.
Nonmonetary Pricing
In the previous section, I discussed the intrinsic value of utility tokens. In the real world, however, we often need to price the token in nonmonetary terms. For example, in a PoW system like Bitcoin, the miners are awarded for their efforts. If it is too easy for the miners to earn new Bitcoin awards, the market price of Bitcoin will never increase since miners will simply sell their tokens for quick profits as soon as they mine them.
To understand this phenomenon in the medium of exchange framework, it is equivalent to adding token supplies to the system at a faster pace than the growth of the network GDP. It causes the token price to fall due to oversupply.
However, most blockchain networks have mechanisms for future users to earn tokens to balance the increasing network GDP. A primary objective for a public token sale is to jump-start the network effect. Once the token sale is complete, it is often desirable to provide ongoing incentives for new users to earn tokens by contributing to the network effect in other ways. All these lead to increased network GDP and an increased token price for everyone.
For protocol designers, it is therefore important to balance the interests of token purchasers and token earners to make sure everyone receives tokens at a fair cost. To see a well-designed protocol, let’s look no further than Bitcoin.
Bitcoin miners do not receive their BTC for free. They need to spend money on electricity, mining machines, data centers, and real estate. If the BTC price goes up in the market, more people will want to join or expand mining operations, hence increasing the computing power (hash rate) on the network. The protocol is designed so that an increased network hash rate will cause the Bitcoin mining algorithm to automatically increase its difficulty, resulting in higher costs for miners. The automatic adjustment of mining difficulty creates a negative feedback cycle that keeps the mining cost always in sync with the current BTC market price. Because of that, there is no free ride in the Bitcoin ecosystem.
In many newer blockchain networks, there are “proof of XYZ” mechanisms for users to earn new tokens from the network. The protocol must have a negative feedback loop similar to the Bitcoin system to ensure that the “cost” of those newly mined or minted tokens is inline with the current market price.
Stable Coins
An important goal of crypto token design is to create a SC that has a stable exchange rate against a fiat currency. There are at least two important use cases for the SC.
A token must have a stable value to be used as a payment utility. No one will use a wildly fluctuating token to purchase goods since you can never be sure what the price will be in the next hour.
An SC can act as a hedge or safe harbor for token traders. As we have seen, crypto tokens’ prices are all highly correlated. They often go up and down at the same time. For a trader, during a down market, there is no safe token where she can park her money to wait for the market bottom. The only choice seems to be exiting and re-entering the market using fiat currencies, creating problems with taxes, trading speed, or algorithm automations. An SC is in great demand in exchanges.
An SC is a poor candidate for a token sale since no one wants to buy a token that does not appreciate. But algorithmic SCs typically come in pairs, and the token that represents the asset pool does have the potential to go up in price. In the next two sections, I will discuss two common approaches to issue SCs on blockchain networks without central banks.
Fully Collateralized Stable Coins
A collateralized SC can always be redeemed for a stable value at any time. For example, users can purchase SCs for $1 each from the issuer. The issuer could be any centralized entity, and it holds the USD as reserve. The issuer promises to redeem and burn SCs at $1 each at any time. The SCs are then used in transactions in a blockchain network. Each transaction will generate a fee collected by the issuer.
A user can choose to hold and reuse SCs, as opposed to exchanging SCs for USDs immediately before and after use. That is because
Redeeming SC for USD requires KYC and banking fees.
Receiving the potential dividend from transaction fees.
The issuer goes through the trouble of holding and exchanging USDs because it could profit from the following:
Potential fees for SC transactions
Interest rate on the USDs it holds as collateral
In this scenario, we have a generic SC issuer. In reality, many entities in a blockchain network have additional incentives to issue SCs and more ways to profit from SCs. Here are some examples:
Cash flow issuers: SCs could be issued by a business that has a stable USD cash flow. SCs could be issued from cash income generated by the business and then be rebated to the customers. This allows the business to subsidize and promote certain user behaviors while creating an SC as a side effect.
Loan issuers: SCs can be created from a crypto loan. The issuer could put up Bitcoins as collateral and issue SCs to spend as USDs. When the SC holders redeem their SCs, the issuer would sell Bitcoins to cover USDs. In this case, there must be rules in place to liquidate the Bitcoin collateral into USD when the Bitcoin price drops.
Payor issuers: SCs could be issued by users with high transaction volumes on a trading network. The user deposits USDs required for payments and issues himself SCs. The user then uses the generated SCs for actual payments. The payment recipients can further reuse the SCs for other payments. The issuer holds the USD deposits as liabilities for the outstanding SCs. In this case, the issuer essentially guarantees payments on the entire network, which is arguably the responsibility of heavy users of the network. A key benefit of this approach is that it can be decentralized.
As we can see, there are many reasons for entities to put up collateral assets and issue SCs. We envision a future world with many different SCs for different application scenarios and issuer benefits.
Algorithmic Stable Coins
Another interesting yet unproven class of SCs is called algorithmic stable coins. The basic idea is to create an asset pool that buys the token when its price drops and sells it when the price rises. That is to use the market mechanism to create an SC. The key benefit for algorithmic SCs is that they do not require full collateral to remain stable. They have the ability to dynamically attract capital and assets into the system to function as collateral as the market fluctuates.
There have been several attempts to create an algorithmic SC. For example, the Maker protocol creates two interlinked tokens. Both tokens are free-fluctuating. But through the trading of Maker MKR token, the asset-backed DAI token is supposed to reach a stable price. I will not go into the details of the Maker protocol in this book. Interested readers can read its white paper at https://makerdao.com/whitepaper/.
Conclusion
In this chapter, I explained common economic designs of blockchain cryptocurrencies. In the rest of the book, we will explore how to actually build blockchain networks with cryptocurrency support.