Further Binance-Peg Token Problems
In previous work we established that the Binance-Peg tokens have had management issues for quite some time. Bloomberg News got Binance to confirm the analysis was correct. Bloomberg quotes a Binance spokesperson as saying:
Recently, the process has been much improved with enhanced discrepancy checks to ensure it’s always backed 1–1.
You can probably guess where this is going. By digging deeper into the process we’ve found both:
- Widespread, and sometimes more dramatic, problems beyond BUSD.
- Inconsistencies and confusing co-mingling of funds up to today.
Here we are going to look into 3 major stablecoins (USDC, USDT and TUSD) in their Binance-Peg form. And then we are going to survey a range of other assets which, to this day, exhibit troubling patterns.
Binance-Peg USDC, USDT and TUSD are supposed to be managed just like BUSD with reserves shown here.
Analysis of Binance-Peg USDC is straightforward. If we compare the balance in the ETH Binance-Peg wallet against the market cap we get:
The run up in early 2021 resembles BUSD: tokens are printed on Binance Smart Chain well before they are deposited into the peg address. But, eventually, the balance catches up.
And then in late 2022 something new happens. In this transaction on 17-Aug-2022 Binance entirely drained the peg wallet, transfering out $1.779 billion. Over the next few months $1 billion Binance-Peg USDC were burned. And then, in this transaction on 6-Dec-2022 about $883 million are sent back in, matching the then-current market cap.
Someone received a loan of something like $1B for about 100 days. It is not clearly exactly what happened…but this is very large, very obviously manual and very recent.
With TUSD we see something similar. The amounts are smaller but the pattern is familar:
This one was undercollateralized for a long while. Then it was overcollateralized for a bit. As of now, per Binance’s own disclosures, it remains overcollateralized.
Also the linked backing wallet for the Binance-Peg is this one. That’s an exchange wallet and not the dedicated Binance-Peg wallet. Are these customer funds? or Peg-backing funds? Or a bit of each?
There is a Binance-Peg USDT. And, per Binance’s current disclosures, it is backed by both an ETH wallet — the Binance-Peg one — and a TRON wallet. As something like half of USDT is on TRON this is fine.
The TRON wallet is here. As you can see there is a recent large transfer of just under $2B into this wallet on 6-Dec-2022 (note: same date as USDC above). That transfer comes from a wallet labelled Binance-Cold here.
This is more confusing. There are 3 known potential sources for Binance-Peg USDT backing: the ETH peg wallet, the TRON peg wallet and the TRON cold wallet. Let’s see how those balances stack up against Binance-Peg USDT market cap:
Note a few things here. First, there are a few times where the yellow line — the market cap — exceeds even the total balance of all 3 accounts. And second, this again looks like a clear co-mingling of funds. The cold wallet that, at least sometimes, backs this token contained over $10B of USDT at peak. That is wildly higher than even the peak market cap of Binance-Peg USDT and can only represent exchange deposits.
Now let’s dig in to the divot in May 2022 where the exchange balances drop but the market cap does not. Between the 7th and 12th of May 2022, during the LUNA incident, $900 million of these B-USDT are printed. And, simultaneously, the cold wallet on TRON drops from $10.5 billion to $86 million. That is effected with these three giant transfers into a wallet labelled “Binance-Hot.” So exchange deposits then. And at the same time the B-USDT market cap reaches nearly $5 billion while the total contents of these 3 wallets drops to $3.8 billion.
The cold wallet transfers look like withdrawals. That’s totally expected around LUNA. But the $900 million unbacked inflow? That is suspicious. And note that we say “unbacked inflow” because there is, visibly, no backing. Binance is welcome to clear this up.
There are three key pieces of information to take away from this section:
- Unbacked, or at least not-properly-backed, tokens seem to be a common feature of the Binance-Peg ecosystem which extends even to tokens like USDC that are very clearly outside Binance’s control.
- There is some degree of co-mingling of funds between peg backing and exchange deposits.
- During the LUNA incident Binance saw massive outflows…but also printed Binance-Peg stabelcoins in large size. These were almost surely unbacked and Binance-Peg as a whole looks to have been significantly undercollateralized.
Binance-Peg DAI looks surprisingly overcollateralized:
And the linked wallet is Binance 8, not the peg wallet. Is that client money? Is it a mix? Let’s just assume, for the sake of argument, that the $20-odd million extra DAI in that wallet are client funds and everything is solvent. WTF bother with a peg wallet if you are not going to use it? Again this is a current problem.
The Paxos stablecoin is almost comically overcollateralized. Again with the contents of what looks like an exchange wallet.
These problems extend beyond stablecoins as well. Tezos’ native token is nearly 2x overcollateralized here:
But that Tezos address is not segregated at all. It’s a Binance staking address on Tezos:
There is nothing like segregation here. And this deposit is subject to Tezos slashing: if the Binance Baker has a problem the balance may be reduced. This also suggests there is some co-mingling with their Earn program. Cause that is what is supposed to end up in the staking wallets:
Aave looks correctly collateralized:
But notice that is again the Binance 8 exchange address:
If we check the peg wallet:
It was funded for 400k by the same wallet. And it still has a 400k token balance. The B-Token page shows 100k BEP2 tokens and 400k BEP20 tokens. The BEP20 tokens were properly backed by the Peg wallet. It looks as though someone got confused, noticed the 100k were unfunded, and then mistakenly sent 500k tokens.
The peg wallet’s tokens were maybe forgotten? It would make much more sense to send 100k more tokens in. But no. The co-mingling looks to cost real money with these tokens, at least, backed in two places.
This is all a proper mess. We have identified, with on-chain evidence, at least the following problems:
- Large, often 10-figure, problems with the backing of Binance-Peg tokens beyond simply somewhat-slow transfers.
- There is obvious mixing of client and peg-backing funds. This raises the possibility that client funds are being used to show the solvency of pegs. We have no direct proof that is happening but this evidence is suspicious to say the least.
- Peg funds are, sometimes, subject to slashing.
- Overall absurd collateral management practices. There is a peg wallet but it is often not used. Certain tokens are overcollateralized to a degree that makes no sense at all. And in at least one case it looks as if funds were forgotten.
Given the obvious lack of effective controls — which, to be clear, Binance has admitted to Bloomberg — this raises real questions around the safety of user funds.
Where did those 500k Aave come from if it wasn’t client deposits? If someone can misplace the initial 400k how can we believe other, more catastrophic, mistakes won’t happen? What happened in May 2022 when unbacked stablecoins got printed at the same time a formerly-$10B+ cold wallet was drained to near 0 with transfers to an exchange hot wallet?
Finally it is worth noting that these questions, and this line of inquiry as a whole, are not available in traditional finance. Completely different tools would be employed and questions asked. This is an exercise in using blockchain transparency. And shows the sort of scrutiny folks need to prepare for.