Introduction

Blockchain technology has rewritten the fundamentals of traditional finance. Through peer-to-peer nodes and distributed ledgers based on cryptography, a financial system that is automated, secure, permissionless, and decentralized has been created. Users can use blockchain networks to transfer funds. And network nodes help verify these transactions, create new blocks, maintain network security, and charge fees.
Although the blockchain makes financial services cheaper and more efficient, with the increasing popularity of smart contracts, some new problems have surfaced. One of the new problems is maximum extractable value (MEV), which refers to network nodes including, excluding, and changing the order of transactions in a block to get the most block reward and transaction fee possible that exceed the normal amount from the block they are responsible for.
Due to MEV, unknowing users may suffer losses due to transactions being manipulated, making the network unreliable. However, the impact of MEV on blockchain networks is not entirely negative. In order to maximize revenue, competition among network nodes also accelerates the development of blockchain technology.

What Is Miner Extractable Value (MEV) ？

Miners include or reorder transactions in the blocks they create for more profits.
The origin of the term Maximum Extractable Value (MEV) comes from another term - Miner Extractable Value (also MEV), which refers to the extra revenue miners can obtain by arbitrarily including, excluding, or reordering transactions in the blocks they create. The earliest research on MEV comes from a study titled Flash Boys 2.0: Frontrunning, Transaction Reordering, and Consensus Instability in Decentralized Exchanges. When studying PoW blockchain networks (e.g. Bitcoin and Ethereum), they found that miners would manipulate block creation and transaction ordering in a way that would allow them to obtain higher-than-standard block rewards and transaction fees.
The arbitrage method of packing and reordering transactions exists across consensus mechanisms.
This phenomenon is not limited to proof-of-work blockchains. Also on proof-of-stake (PoS) blockchains, validator nodes can earn additional income from users when creating new blocks, packing pending transactions, and reordering transactions. The term “miner extractable value” ended up being replaced by the broader “maximum extractable value”, which is a form of arbitrage that is risk-free or low-risk.

How MEV Came into Being

The sustainability of a blockchain network is based on financial incentives. Network nodes must continuously be rewarded so that they would be incentivized to provide safe and reliable services. Due to the network latency in different regions and the fact that distributed ledgers require time to reach a consensus when a user sends a transaction on the blockchain, the transaction will not be processed immediately but will be stored in a memory pool (also known as a mempool). When a new block is generated, the unprocessed transactions in the mempool will be added to the new block, and transaction data will be recorded in the ledger.
There are differences in the information and synchronization time among nodes, which allow the order of transactions to be changed before a new block is generated.
Because of the mempool, nodes can share information about certain transactions, evaluate the current network traffic and reorder transactions. Nodes are incentivized by money. So when the mempool is full and cannot accommodate more pending transactions, nodes will first process transactions that pay more fees. It will take longer for transactions paying lower fees to be packed into a block. In some cases, they might even get moved out of the mempool which will result in transaction failure.

When multiple users pay different transaction fees, miners (or nodes) will pack the transactions from the mempool in order, according to how high the fees are. (Source: ChainLink blog)

Since the information of pending transactions in the mempool is public, and nodes have the rights to create new blocks and order transactions, they can easily order the transactions to extract the most value by adding or removing transactions. For example, when a node responsible for packing and accounting notices that a user has requested a transaction to buy 1 ETH with USDT, the node can buy 1 ETH before this user does and then sell it out at a higher price to complete the arbitrage.
Nodes can sell accounting and transaction ordering rights to other arbitrageurs.
Nodes do not have to manipulate transactions themselves to obtain additional revenue. They can also sell their accounting and transaction ordering rights, so that if other users identify arbitrage opportunities, they may be willing to pay higher fees to these nodes to have their transactions smoothly packed into blocks. Either way, nodes’ centralized accounting and transaction ordering rights can be used to take money out of users’ pockets.
The emergence of MEV is not only because pending transactions in the memory pool can be manipulated by nodes, but also the economic inefficiency that gives space to arbitrage. For example, a user buying or selling a large amount of value on a decentralized exchange (DEX) will cause price slippage. Then nodes and arbitrage bots can borrow from liquidity pools of other decentralized exchanges to exploit the price difference.
Another key factor that makes MEV possible is the transaction simulation algorithms. Since the blockchain is a transparent public ledger, the liquidity and market depth of all on-chain assets are public to everyone. The impact of each trade on the market price can be predicted and then arbitrageurs can work out the correct size of their positions and the transaction order to extract value.

Common MEV Cases

Blockchain network nodes can harvest MEV in multiple ways. The following are a few common examples.

1. DEX Arbitrage
   Arbitrage between DEXes is the simplest and most intuitive form of MEV. If the algorithmic quotation of an asset in DEX A is 100 USDT, while the algorithmic quotation in DEX B is 110 USDT, whoever spots the price difference can buy the asset from DEX A, and then sell it in DEX B to exploit the price difference. This type of arbitrage also exists in traditional financial markets. Since it is basically risk-free, the competition for such arbitrage opportunities is quite fierce.

2. Liquidation in Lending Protocols
   In lending protocols that require collateral (such as AAVE and Maker), users can deposit crypto assets (e.g. ETH) to lend other crypto assets (e.g. USDT). As long as the value of the borrower’s collateral is sufficient, the borrower can continuously lend the cryptos from the protocol.
   Each protocol has its own standards for collaterals. For example, when the value of the collateral becomes lower than 70% of the value of the borrowed cryptocurrencies due to price falls, in case the price continues to fall and the borrower becomes insolvent, the lending protocol can sell the collateral to repay the loan, which is known as liquidation.
   Lending protocols allow all borrowers to liquidate the collateral to pay off the loan immediately. When liquidation happens, the borrower has to pay a high liquidation fee, a portion of which goes to the liquidator. Performing liquidation is also a risk-free source of income. There are many bots designed to constantly look for borrowers who will be liquidated, preemptively propose liquidation, and collect liquidation fees.

3. Sandwich Attack
   Sandwich attack is another common MEV extraction method, which is to insert transactions before and after a target transaction so that the victim trader whose transaction gets “sandwiched” will have to pay more.
   For example, a user wants to use USDT to buy 10,000 ETH on Uniswap. Since this is a very large amount of value, the price of ETH in the liquidity pool will rise (P1 → P2 in the graph below).
   After an arbitrage bot “sniffs” out the pending transaction, it can front-run the victim and buy 10,000 ETH in the market, so that the market price will reach P2. When the victim’s transaction is executed, the price of ETH will be further pushed up from P2 to P3.
   Now the arbitrage bot will sell the previously purchased 10,000 ETH, making the price return to P2 again. As a result, the cost of purchasing ETH has increased for the victim, and the extra cost becomes the arbitrage bot’s profit.

Arbitrage bots can sandwich users’ transactions to profit from the price difference.

1. Other Arbitrage Opportunities
   All three arbitrage methods listed above are the most common ones. Therefore, the competition for new participants to obtain profits is quite fierce.
   However, in some new markets (e.g. NFT), it is still possible to make the blockchain network execute your own transactions (e.g. batch minting NFTs) through front-running, or to batch buy or sell when price changes are detected. However, the liquidity in new markets is usually very low, so arbitrage there would come with high risks.
   Another way to earn profits is by introducing a large amount of liquidity at once. For example, in Uniswap V3, liquidity providers have control over what price ranges their capital is allocated to and can therefore collect higher transaction fees. In extreme cases, arbitrageurs add liquidity every time they detect other users’ transactions and then redeem liquidity immediately after the transaction is completed. This way, arbitrageurs monopolize nearly 100% of the transaction fees, eliminating the risk of long-term impermanent losses.

Pros & Cons of MEV

Although on the surface, MEV only increases the revenue of network nodes at the expense of users, similar to how nodes charge additional taxes from network users. It is difficult to judge if this value reallocation is good or bad.
For network nodes and arbitrageurs, being able to change the order of transactions and harvest MEV is definitely a good thing. These arbitrage opportunities will also attract more people to join, which will improve the security and decentralization of the blockchain network.
However, for users, the cost of on-chain transactions increases and arbitrageur’s front-running transactions occupy the bandwidth of the network, leading to bad user experience. On the other hand, these hard-working arbitrageurs also improve the efficiency of the blockchain, so that the cryptocurrency assets on the chain will not have too large a price difference. And the execution of the liquidation also ensures the safety of using lending protocols.
From the perspective of the blockchain network, MEV does cause some problems. Those top arbitrageur nodes that are good at extracting value from transactions will accumulate more and more resources, making the blockchain network more centralized. In extreme cases, arbitrageur nodes may even try to tamper with old blocks for extra bucks.
To conclude, the impact of MEV has multiple dimensions, but there is no doubt that the research on MEV is a good opportunity to find out how blockchain technologies could be better.

How to Improve MEV？

The foundation of the MEV is the transparency of mempools and nodes’ rights to freely arrange the order of transactions. Therefore, if the mempool is no longer transparent or if nodes are forced to pack transactions in a certain order, additional fees can be avoided for users.
For example, Automata Network uses an algorithm called Conveyor, which draws pending transactions from the mempool in a certain order and attaches Automata’s node signature to the transactions. Transactions that are not added by Conveyor will be detected. Only the ones processed by Conveyor can be packed to new blocks.

Source: Automata Blog
Chainlink uses an algorithm called Fair Sequencing Services (FSS) to address the issue of MEV. FSS requires smart contracts to sequence transactions according to certain parameters, such as time (when the transaction enters the memory pool), transaction fee, transaction amount, and transaction type.
FSS can also encrypt the sequence of transactions, which can only be decrypted after it has been submitted. This encryption is also a form of Proposer/Builder Separation.

Source: Chainlink Blog
There is also a practice called MEVA (short for MEV Auctions). MEVA recognizes the existence of MEV, and also encourages all nodes and arbitrageurs to obtain MAV in any form as much as possible. Part of the proceeds of the auctions will be transferred to public funds and given back to all users of the protocol.
Optimism is a layer 2 protocol that uses the MEVA model. To some extent, MEVA makes users pay “tax” to boost the development of the protocol because nodes and arbitrageurs can be incentivized to improve the efficiency of the network.
Using on-chain aggregator protocols helps reduce the negative impact of MEV and improve capital efficiency.
Aggregator protocols can also shield users from the negative impact of MEV. The price difference of the same cryptocurrency asset across different liquidity pools often comes from users’ large amounts of buying or selling on one DEX. In theory, users can split large transactions into different liquidity pools to reduce their costs, even though doing so will take time and patience. Aggregators like 1inch will work out the best way for users to conduct large transactions at one time, enhancing capital efficiency.
It is not true that transactions processed by aggregators cannot be exploited by network nodes anymore. With more aggregators being invented and related technologies getting more advanced, there will be less space for arbitrage. These aggregator protocols can actually be seen as “cheaper arbitrageurs”. MEV will still continue to exist. If aggregators can be cheaper, then users have no reason to pay more to other arbitrageurs.

The Impact of the Merge on MEV

Under PoW, miners can maximize their profits through priority transaction ordering, enabling DEX arbitrage, sandwich attacks, etc.
After Ethereum transitions to PoS, the daily output of ETH will be greatly reduced. The daily block reward will be reduced by 90% from 14,600 ETH to about 1,600. Under PoS, whether you stake 32 ETH or 100 ETH, the reward for verifying each block is fixed.
Therefore, MEV will be very important for PoS validators to get more rewards. MEV will also encourage people to stake ether and become validators. Therefore, MEV will be very important for Ethereum, which values ​​decentralization and security.

Annual validator reward is the sum of staking reward and MEV. (Source: Flashbots)

Both PoW and PoS face some problems. Exploring MEV benefits people with higher capital more. Because miners (validators) do not necessarily have to execute transactions themselves to harvest MEV, they can act as proposers and accept offers from block builders to give certain transactions priority to get verified.
Therefore, validators with high capital can pay expensive fees to allow transactions that benefit them to be preferentially processed. This means these elite nodes can, directly and indirectly, control the transaction data of the block. If this problem cannot be solved, Ethereum will have less decentralization.
After the Merge, most validators will deploy MEV-boost, a free and open-source software built by Flashbots together with Ethereum developers and researchers. MEV-boost helps Ethereum lessen the negative impact of MEV and make the harvest of MEV more democratic.
MEV-boost is a solution based on PBS (Proposer Builder Separation), which can solve the problem of centralization of nodes after it is implemented on Ethereum.

What is PBS (Proposer/Builder Separation)
Proposer/Builder Separation is to split the block construction role from the block proposal role of a node. In this way, nodes cannot randomly add, remove or change the order of transactions in their own interests. Components of the PBS are:

1. Block proposers: They propose blocks and send the distributed ledger, also known as validators.
2. Block builders: They try to build the most valuable block, and maximize the value of a block by searching MEV themselves or by accepting bundles from MEV searchers.
3. MEV searchers: They try to identify profitable transactions that are pending reviewing in the public mempool in order to get potential arbitrage opportunities. For example, if a MEV searcher finds a large transaction in the mempool, he/she may conduct a sandwich attack.
4. Bundles: MEV searchers produce bundles of a single or multiple transactions. The bundle must be placed at the top of the block and cannot be split.
   For example, most of the PoS investors on Ethereum currently use Lido or other liquid staking protocols, so the number of builders are much smaller than that of proposers. On this occasion, nodes and validators of Ethereum become overly centralized. However, PBS can effectively solve this problem, because block building and block proposing are assigned to different roles in the network.
   The builder can extract MEV or submit bids to the proposer after accepting the bundles provided by the MEV searcher. The proposer must approve the bundle from the builder. The proposer does not know the contents of the bundle, which represents an effective way to prevent the proposer from controlling the transaction order and stealing MEV.

Why run MEV-boost
MEV-boost is a solution to PBS, which allows individual stakers to participate in a blockchain network and increase its decentralization.

1. Users and searchers send transactions to block builders through the public p2p txpool or through direct channels.
2. Builders construct execution payloads using these transactions and header parameters provided by validators. Builders may directly set the validator’s fee recipient address.
3. Relays receive execution payloads from builders and verify the validity of the payloads as well as calculate the payload value (amount of ETH paid to the fee recipient).
4. Escrows receive the full execution payloads from relays to provide data availability.
5. Validators receive execution payload headers from relays (execution payloads stripped of the transaction content). The validator selects the most valuable header, signs the payload, and returns it to the relay and escrow to be propagated to the network.
   
   Block construction process for mev-boost, referenced from the Manifold Finance blog

This proposed architecture allows validators to outsource the task of block construction to a network of third party block builders. While the validators have the ability to include any payload into the chain, the network neutrality and validator revenues are maximized when the validator’s job is limited to selecting the payload which pays them the most ETH.

Conclusion

In terms of decentralization and fairness, MEV does have advantages, despite its many disadvantages.
Generally speaking, the blockchain promises to create a fair, permissionless, trustless and decentralized financial system for users. However, if we take a closer look at it, we will find that there are still trusted third-party intermediaries in the blockchain space. In traditional finance, the intermediaries are government organs and banks, while in blockchain networks, they are miners and validators who are responsible for creating new blocks and manipulating the order of transactions. They utilize their power to add, remove or rearrange transactions in blocks to get addition al profits from users.
Economic incentives drive more nodes to compete, hence enhancing decentralization.
Although no user would be happy to pay this unnecessary network usage tax, or to accept the congested blockchain network caused by arbitrage bots sending a large number of junk transactions, the existence of MEV brings advantages in addition to disadvantages. Economic incentives drive more nodes to join, making the blockchain network more decentralized and ensuring the stability of price in different markets. Users can trade at any time without the need to compare prices different markets beforehand. How to maximize its advantages while minimizing the disadvantages is a key issue to discuss in the future.

Solutions to mitigate the impact of MEV have been proposed one after another, but the balance of multi-role power in the market still needs to be considered.
As for how to mitigate or even eliminate MEV, many different solutions have been proposed, such as fair sequencing services, encrypted privacy transactions, block ordering rights auction, proposer/builder separation, etc. These methods are novel, and are still inconclusive whether they can protect the interests of both miner and arbitrageurs and ensure the steady growth of the blockchain network. Each transaction has values associated with it. Perhaps it is acceptable to allow arbitrageurs to extract some of the value within reasonable limits.
MEV will continue to exist, and a free market will eventually achieve a dynamic equilibrium.
As blockchain technology matures, it is expected that MEV strategies with more complex, comprehensive and integrated multi-blockchain networks will emerge. Some are concerned this situation will make the operators of nodes or arbitrage bots monopolize resources and cause a crisis of centralization, but the fact is that the expansion of knowledge will inevitably bring competitors. In the long run, MEV is a market full of fierce competition, in which inefficient nodes and blockchain networks will be eliminated, while those with the best performances will be supported by users. In the race for resources, the balance between efficiency and decentralization will be achieved. MEV may continue to exist but will become cheaper and more competitive, acting as a key driver for the growth of the crypto ecosystem.