Where has all the money gone indeed?
This is a response to Frances Coppola’s piece of a similar name. While generally agreeing with that analysis one part sticks out:
The overall picture that emerges from this, together with the extraordinary series of loan agreements in the bundles, is of a company that was massively over-leveraged and had for some time been experiencing cash flow strain. It borrowed heavily to ride the wave of crypto appreciation during the pandemic. “The money will never run out” was its philosophy. But when the market turned and crypto prices started to fall, it had to borrow even more to maintain its collateral and meet its obligations.
Hyman Minsky called borrowing to obtain cash to meet existing obligations “Ponzi borrowing”, because it depends on there being a constant supply of new lenders. This is ultimately unsustainable, of course, but it can be maintained for quite a while if the lenders don’t know what is going on and there are no shocks to the system. And indeed, 3AC managed to conceal the scale of its indebtedness from lenders, customers and markets alike. But as the crypto winter deepened, it found it harder and harder to raise funds. Lenders demanded more collateral than it could afford to provide and better information than it wanted to give. One lender pulled out of a loan deal when 3AC refused to provide a balance sheet.
This seems too kind, at least to some people. Let’s take another look at the cash flow statement linked there:
Every dollar borrowed was ploughed into “purchases of digital assets.” Given the quantum of lending we know was going on at the end it may not be fair to suggest — for some bilateral relationships at least — that anything was concealed.
If you add up those five marked numbers in the table you get 0. They borrowed cash and spent all of it on digital assets. They were not borrowing to pay debt service or rent an office or (at that point) buy a yacht. They were borrowing to accumulate digital assets of various sorts. Net cash was 0.
The Lenders
Now let’s try organizing the liabilities again but this time by date. And we will ignore the gigantic Genesis loans as they was clearly ongoing, at some level, the whole time. The agreement is dated 10 Jan 2019 after all. The remaining gross borrowings total something like $1 billion.
There is a reasonably solid argument those blue bars on the left knew exactly what they were doing. If we log-scale the loan quantities it is pretty clear these are all same order of magnitude:
This is a classic regime change situation. On the left the lending is to arb this spread. There is then a gap in the middle when the spread goes away. And then a flurry of borrowing nearer the end to finance losses.
It is impossible to tell from here what messages each group got. But there are two separate groups of lenders here: people who financed a risky arb trade and people who financed a blown-up SIV.
There was widespread reporting of people doing GBTC premium trades like this. And 3AC’s own 13G came out in early June 2020. As we have previously reported Shu Zu was tweeting about this thing since before the chart starts over there on the left.
One meaningfully-sized loan has no date given: the $65 million FalconX loan. But that company only raised significant assets in Sep 2021 and then Jun 2022 so they are almost certainly on the later side of that gap as well. They claim to be all fine and significantly overcollateralized with stuff-that-is-not-GBTC in this piece. So good on them. Perhaps once we see those docs it can provide a template of how to manage some of this with less damage. Though, honestly, it is difficult to puzzle out what that was and we very much look forward to their filings.
Post-LUNA Lending
There are only two post-LUNA loans with firm dates: CoinList and Mirana. Those were almost certainly lending into a hole. For these loans we agree 100% with Frances’ assessment.
For the others, yeah, the jury is still out. There is a huge difference between financing this thing with a pile of BTC you already had in 2019 and rolling it for 2 years vs just lending USD to 3AC near the end. And these charts tell that story very clearly.
