By Karthika Swamy Cohen
Bitcoin’s value more than tripled in the second quarter of 2017, making it the most prominent cryptocurrency on the market. In a minisymposium talk titled “Monopoly Without a Monopolist: An Economic Analysis of the Bitcoin Payment System” at the SIAM Annual Meeting being held in Portland, Ore., this week, Ciamac Moallemi of Columbia University overviewed the economics of the system, seeking to give answers to basic questions such as, how does the system pay for its infrastructure? How does it generate revenue? What determines usage fees?
Moallemi delved into Bitcoin from the perspective of a payment system rather than a speculative currency, and focused on evaluating transaction costs of the system endogenously.
He began by contrasting Bitcoin to traditional systems like PayPal, Venmo or a bank wire transfer. “These are centralized systems – with a central entity operating the system,” he said. “They decide what the rules are, how much infrastructure to provision, and what the fees are. They provide trust. Why do I trust Venmo? If they take my money, I will sue them.”
The abovementioned systems are natural monopolies. Bitcoin, on the other hand, is a decentralized platform. There is no central entity operating the system, and it lacks a controlling owner who can set fees. It comprises two main constituencies: users, who make and receive payments, and profit-seekers or “miners,” who maintain the system's infrastructure. The latter make up the decentralized service providers. Users make payments via transaction fees and miners monitor the system to ensure there is no fraud or criminal activity.
“How is it decided how many miners there will be and what the charges will be?” Moallemi asked. “It's radically different and amazing that the Bitcoin system can transact huge amounts of money without an operational entity and legal system.”
He proceeded to answer this by describing how the blockchain system of Bitcoin works, focusing on marketing design and the economic perspective.
Bitcoin does away with monopoly pricing, where a single entity controls the system. On the flip side, the decentralized fashion in which the system is run leads to congestion and delays.
The blockchain protocol sets the system rules, tracks balances, and ensures that the system is trusted and properly funded. An atomic unit of block chain is a list of all past Bitcoin transactions organized in sets of blocks, or ledgers of transaction data.
A user who initiates the transaction decides the transaction fee. This evolves in time as more transactions are made.
Anyone can be a miner or profit-seeker – it’s a free entry/exit system. All miners can verify payments and account balances. The pool of transactions is transmitted to all miners, which they can all review. Approximately every ten minutes, one miner is randomly picked to select the next block of transactions, verify, and assemble them. This block is then transmitted to other miners. Block sizes are limited.
Each miner checks that all transactions are valid and adds it to his or her local copy of block chain.
“If someone cheats there is no legal system to counter that, but you have many miners checking each other's work,” Moallemi explained. “Other miners will notice and reject that block.”
So why do miners bother to do this? There are rewards for mining a block.
One type of reward for mining a block—called minting coins—is 12.5 bitcoins. The second mechanism of getting rewards is the transaction fee – a miner gets to keep all transactions in blocks that he or she mines. Currently, the infrastructure is funded via both minted coins and user transaction fees. The new coins, however are slated to phase out, which means that in a few years, transaction fees will remain the sole revenue source.
The total system revenue includes the payments made to miners, which equals the total transaction fees paid by users.
The number of miners determines the infrastructure level and the amount of security. The more miners there are, the costlier it is to cheat. However, the throughput is independent of the number of miners. As far as the protocol is concerned, the selection of a miner for the next block requires solving a computational challenge and there is a cost to that.
Miners are the price takers. While they do not have market power to set transactions or affect the behavior of users, they can maximize rewards by choosing the highest value transactions up to block capacity.
“From the perspective of miners, the analysis is actually pretty easy,” said Moallemi. The easiest way for a miner to get value is by processing the highest value transaction fees.
Analysis on the user side to determine transaction fees presents a more interesting challenge. Since the user knows that miners will process transactions based on the value of transaction fees, the fees they pay can be viewed as a bid for prioritization in the system. For a user with zero value for time, the bid will equal zero.
The delay in the system depends on what one bids, but also on what others bid – in other words, for a given user, delay is contingent upon the population of people who bid higher than him or her.
The decentralized Bitcoin system can generate revenue without excluding any transactions. A monopoly system, in contrast, will process transactions without delay but sets a minimal fee, which may exclude some potential users.
In the Bitcoin system, if there is no congestion and no delay, meaningful transaction fees will disappear, leading to collapse of the system in the long run. Users will pay negligent transaction fees without substantial congestion, which will result in loss of revenue to fund the network of computers that implement digital currency transactions.
Moallemi thus demonstrated that while the Bitcoin system eliminates dead-weight loss from monopoly, delays form an essential feature that generate revenue. The significant congestion allows the system’s blockchain technology to raise the necessary revenue to fund its computer infrastructure. Additionally, the processing of large but infrequent blocks of transactions in the current system design makes it less efficient at raising revenue.
Moallemi and his colleagues propose several alternatives to ensure stability and profitability, such as, creation of smaller but more frequent transaction block sizes, and dynamic adjustment of network throughput to maintain a target congestion level.