Bank of Canada assesses flash loans relevance in latest study

The Bank of Canada published a staff discussion paper on March 21, analyzing flash loans and their relevance for policymakers, as well as potential risks.

The study introduced flash loans as blockchain-native financial tools that enable users to borrow crypto without posting collateral, provided the loan is repaid within a single atomic transaction.

What is notable about the publication is that a staff discussion paper represents a completed staff study on subjects deemed relevant to the central bank and is part of the Bank of Canada’s broader mandate to evaluate emerging technologies pertinent to financial stability and market structure.

Broad relevance

One of the study’s main takeaways was finding the broader relevance of flash loans for policymakers.

Jack Mandin, the study’s author and a former Bank of Canada research assistant, noted that while flash loans are currently confined to blockchain networks, the underlying concept could be extended to tokenized financial infrastructure if similar technical conditions are met.

Such concepts included atomic and risk-free lending, which could result in systems capable of supporting atomic transactions and programmable assets.

The study also raised concerns about financial stability. Direct risks may emerge if financial institutions begin integrating smart contract-based lending.

Additionally, it highlighted that contagion risks are plausible where blockchain-based assets, including those linked to flash loan activity, become embedded in traditional financial products, such as exchange-traded funds.

Comprehensive dataset on flash loan activity

The paper also documented the development and usage of flash loans from their inception in 2018 through early 2025.

Mandin compiled a novel dataset covering nearly 24 million flash loan events and over $3 trillion in total volume across 11 Ethereum Virtual Machine (EVM)-compatible blockchains, including Ethereum, Arbitrum, and Optimism.

The analysis identified trends in flash loan design, usage patterns, and technical implications for DeFi. It also explored three core flash loan models: basic flash loans, flash swaps, and flash mints.

Each design differs in how liquidity is sourced and repaid, with flash mints offering virtually unlimited borrowing capacity through on-demand token issuance and burning.

The study classified flash loan usage into five primary categories. Positive use cases include arbitrage, liquidations, and liquidity management, while negative use cases involve wash trading and smart contract exploits.

Arbitrage operations accounted for over 75% of all flash loan events, indicating a strong link between usage and decentralized market efficiency.

The research also highlighted how flash loans have facilitated known vulnerabilities in DeFi protocols, including price oracle attacks and reentrancy exploits. These issues have led to material financial losses in some instances.

Consequently, although most of the flash loan activity is concentrated in legitimate financial operations, high-value transactions with unclear purposes suggest the likelihood of unreported or undetected exploits.