Thread by @hasufl

Is BitMEX giving some much-needed transparency about their insurance fund?

Thread

https://twitter.com/BitMEXResearch/status/1094894482074542080

First, they compared the size of the fund to safeguards employed by traditional exchanges like the CME. The sum of CME's insurance is around 29% of BitMEX's insurance relative to their respective annual notional trading volumes.

That difference in itself is totally fine, as the CME is a KYC exchange and also has more of an institutional audience. Losses can, in theory, be recouped via the legal system, which is hardly possible on a non-KYC exchange like BitMEX.

However...

The insurance provided by the CME does explicitly NOT live on their own balance sheet. This is my big problem with the BitMEX insurance fund. CME has no incentive to overinsure their customers because they don't get to pocket excess money that is never paid out.

So are BitMEX customers overinsured? The post gives a counterexample: In March 2017, the day the Winklevoss ETF was rejected, Bitcoin's price dropped 30% and depleted the insurance fund, causing ADL to happen. What it doesn't mention: The fund had only 66 BTC in it.

Better examples can be found in 2018:
16.1.2018, the fund lost 132 BTC (5% of total)
12.4.2018, the fund lost 641 BTC (13% of total)
18.7.2018, the fund lost 103 BTC (1% of total)
22.8.2018, the fund lost 155 BTC (1% of total)

"Given the volatility [in the crypto space], it’s not impossible that the fund is drained down to zero again."

I see that as increasingly unlikely, especially considering that since the last considerable loss (in April 2018), the fund has grown from 4.8k to 21k BTC.

So the fund is arguably too big, but that's not my main concern. It is that BitMEX itself owns the fund and not an external insurer or industry consortium. If they see the fund as what it is, an asset on their balance sheet, they are incentivized to grow it as much as possible.

ADL means winners pay for the losers, but it also guarantees that customers will never pay more than is needed. The insurance fund means losers pay for losers, but makes them, in theory, overpay up 10x or more for the same security. The difference goes to BitMEX as profits.

In summary, the post explains well why a high-leverage exchange *needs* some kind of safeguards, and if possible multiple layers of them. It gives no satisfying answers to why the BitMEX model of an internal "losers-pay-for-losers" fund is superior to existing models.

Here are some more questions I would like to see answered:

1) At what size would you start calling the fund overcapitalized? Any plans on capping the fund and changing the model once that size is reached?

2) What do you plan on doing with any excess capital?

3) The fund is segmented per contract, but the segments are not rolled over to the next quarter. How much money is reversed for each contract? What happens to the money when a contract expires?

PS: No matter how you turn it, traditional exchanges don't make money when customers get liquidated. BitMEX does.

This crass misalignment of incentives births all kinds of other speculation: Does BM weaponize server problems? Does it trade against their customers?

Changing their financial security model so they no longer benefit from liquidations would go a long way to shut down these criticisms.