TL;DR - The change to not needing swap until you exceed your trading wallet USD balance will actually reduce risk for lenders, but only if the calculation for liquidation prices remains unchanged (and if lenders are paid before traders in case of liquidation). It will however do this at the cost of reducing the demand for swaps by anywhere between 40% to 100%.BFX, can you also please clarify what effect the changes will have on liquidation prices? I'm going to run through an example to clarify my understanding of all the changes, please correct.
Let's say we have a trader, with the 1000USD and 1BTC in their trading account. BTC is trading for 600 say, so their margin balance is $1600.
They have 2.5:1 leverage switched on, so their liquidation price is when their effective margin reaches 13% of their margin.
Their tradeable balance is (note that swap interest is a negative value):
Pre-July 21 calc:
leverage * margin balance (in USD) + unrealized profit + unrealized swap - balance spent on any open trading position(s)
2.5 * 1600 + 0 + 0 - 0 = 4000
Post-July 21 calc:
leverage * margin balance (in USD) + unrealized profit + unrealized swap - balance spent on any open trading position(s) - value of BTC component of margin balance*
2.5 * 1600 + 0 + 0 - 0 - 600 = 3400
(The BTC value of the margin balance is subtracted from tradable balance because this is treated essentially like taking a BTC 'position' with some of your tradeable balance)
What I call 'effective margin' is margin balance plus unrealized profit plus unrealized swap. At the moment a position is opened it is always = margin balance as there is not yet any profit or swap. When the margin ratio = effective margin/margin balance hits the trigger, (in this case 13%) the trader is liquidated. This is how we calculate liquidation price.
If we assume a simple position of n bitcoins bought at price P
buy and C
USD USD plus C
BTC BTC in the trading account, then Margin ratio equals:
( n*P - n*P
buy + I
swap + C
USD + P*C
BTC) / (C
USD + P*C
BTC)
Liquidation price is when this equals 0.13, so
P
liq = (n*P
buy - 0.87*C
USD - I
swap) / (n + 0.87*C
BTC)
Lets say the trader opens a position by buying 5BTC at $600. Their tradeable balance and liquidation prices at the moment the position is opened look like the following:
Pre-July 21 calculation:
Tradeable balance = 2.5 * 1600 = 4000
Active swaps = 5 * 600 = 3000
Liquidation price = ( 5*600 - 0.87*1000 - 0 )/( 5 + 0.87*1 ) = $362.86
Post-July 21 calculation:
Tradeable balance = 2.5 * 1600 - 600 = 3400
Active swaps = 5 * 600 - 1000 = 2000
Liquidation price = ( 5*600 - 0.87*1000 - 0 )/( 5 + 0.87*1 ) = $362.86
(BFX - Is this new liquidation price calc correct ie it remains the same? Does anything change?)
OK, that's all good, but if the liquidation prices are the same, where does the reduced risk come from? The answer is that this change probably doesn't reduce the risk of a cascading liquidation by much, (assuming that I'm right and the liq. price calc doesn't change) but if it works the way I am assuming then lenders will be more protected because even in the case of a forced liquidation, it will be the trader's money that is put at risk first. Let's see what I mean.
Under the pre July-21 system, a liquidation causing a loss to lenders will occur when the effective margin drops below zero. This is when the loss is equal to the total margin balance, so when
P
losing = (n*P
buy - (C
USD - I
swap) / (n + C
BTC)
This will occur at (5*600 - 1000 - 0)/(5+1) = $333.33.
We can see that liquidating the position by selling the 5 bitcoins, plus the 1 bitcoin collateral, will yield 6 * 333.33 = 2000. This plus the 1000 in the trader's wallet will be just enough to pay back the lender their $3000 swap.
OK, what about after July 21? Say price goes to $333.33 and we liquidate. We get the same $2000, and the trader also has the $1000 in his wallet. The lender gets paid back the $2000, and the trader at least still has his $1000 USD collateral too. There's an 'extra' $1000 USD of margin in play here!
In the case that the liquidation price is below this amount, I would assume that the lender takes priority (again BFX need to clarify here!). So this means that the actual price that will cause a loss to lenders post July 21 is when effective margin = negative USD collateral, so:
P
losing = (n*P
buy - (2*C
USD - I
swap) / (n + C
BTC)
So yes this change will result in reduced risk for lenders, but it will also mean that demand for swaps to cover the portion of the position that otherwise would be bought with the trading wallet USD balance will vanish. Given that for each individual, their current leverage is bounded between 1 (they are borrowing less than their margin balance) and 2.5, this drop in demand is bounded between 100% and 40%. Obviously most people using the margin trading feature have leverage above 1, so it's unlikely that the drop will be near to 100, but it's also unlikely to be near to 40% either. I would guess between 50 - 60%.
It's also worth noting that this part of the change doesn't affect the likelihood of a cascading flash crash very much. However the change in the treatment of BTC/LTC collateral will have an effect there, by reducing the effective leverage possible for these traders.
I think this change by BFX is as much a PR move as it is a move truly intended to avert systemic risk. The headline 'swaps active' number is going to drop by a huge amount; this is all BFX management really want, cos they hope that then people will stop whinging about it in the forums and making them look bad. Given Giancarlo's response to the completely reasonable thread about there being a BFX credit bubble (he called it FUD and was childishly abusive) I think the real motivations here are clear.
Lender beware.