The Unsurprising Surprise of FTX
Another Quarter and another Crypto-crisis. This time FTX halted withdrawals and has filed for bankruptcy as a results of what we’ve coined a “Denial of Credibility” attack. And the next question is, who will be left bearing the resultant losses? Presumably various prominent names, perhaps including some TradFi Hedge Funds. FTX was widely hailed as a relatively safe counterparty and wrote substantial volumes of OTC Futures.
But was this all so unpredictable? Earlier this year, we cautioned that FTX was much riskier than the consensus view. Most of the crypto-firms we told this to were sceptical. They cited the strong tech team, with their deft handling of the Merge, the superb customer service and the substantial equity valuation established by FTX’s VC backers. We argued that this was largely beside the point. In this post we explain how we formed our view and explore the implications for crypto-counterparty risk management.
The table shows the default risk ratings we circulated in July 2022 for ten of the larger exchanges. These ratings were based on publicly available data from various sources, including reported volumes. The identification and weighting of these variables was determined by prior default experience, both within crypto-markets and the wider economy. The aggregate default rate across the 100 rated exchanges was calibrated using observed historical crypto-exchange default rates.
The output was one-year default probabilities. We also mapped these to the equivalent S&P ratings based on S&P’s own published empirical default rates*. One source of cross-validation was our estimate for Coinbase. Because Coinbase has publicly traded securities, these can be used to generate implied default risk estimates. The stock price is used by Moody’s Analytics to publish an Expected Default Frequency based on KMV’s risk ratings**. Likewise, the bonds contain an implied default risk in their yield premium. Both these estimates corroborated our independent analysis.
The first observation from the table is the relatively large size of these default risks. Even Kraken, the safest counterparty shown, has an implied BB- rating. For contrast, Bank of America is rated A+, with an implied one-year default risk of 0.04%***. Kraken is about x50 riskier. Moreover, such TradFi banks enjoy implied government protection. The takeaway is that counterparty risk management is much, much more important in crypto-markets. The firms are substantially riskier and there’s no government backstop.
Finally, what about FTX? How come we concluded, at an 11.3% default probability, that they were so much riskier than the other majors, like Binance? The reality was that FTX actually had a lot of risk indicators. The firm was comparatively young. It had high reported volumes for what was otherwise a relatively small asset base. It was growing fast. None of these factors in themselves lead to default. Yet they make a firm vulnerable. The Denial of Credibility attack that took them down was enabled by that underlying instability.
* Kraemer, N. W. et al (2022, April 13) Default, Transition, and Recovery: 2021 Annual Global Corporate Default And Rating Transition Study www.spglobal.com/ratings/en/research/articles/220413-default-transition-and-recovery-2021-annual-global-corporate-default-and-rating-transition-study-12336975
** Rehm, F. & Rudolf, M. (2000). KMV Credit Risk Modeling. In Frenkel, M., Hommel, U., & Rudolf, M. (Eds) Risk Management. Springer. https://doi.org/10.1007/978-3-662-04008-9_8
*** Available at www.spglobal.com/ratings/en/index with EntityId = 413007
