Serving the Quantitative Finance Community

 
User avatar
bloodynri
Topic Author
Posts: 5
Joined: June 15th, 2011, 12:27 pm

Pricing default risk in cryptos

August 16th, 2018, 1:39 pm

I'm looking to figure out how to price "insurance" against a counter-party defaulting in an OTC cryptocurrency transaction. I think the first measure would be to calculate VaR? I'm planning on modeling bitcoin using a stochastic process and calibrating it using historical data. I'm not quite sure which model to use for bitcoin. I've seen some papers use a GARCH model. 
In general, am I approaching this correctly? Any suggestions?
Thanks! 
 
User avatar
rmax
Posts: 374
Joined: December 8th, 2005, 9:31 am

Re: Pricing default risk in cryptos

August 16th, 2018, 4:16 pm

I'm looking to figure out how to price "insurance" against a counter-party defaulting in an OTC cryptocurrency transaction. I think the first measure would be to calculate VaR? I'm planning on modeling bitcoin using a stochastic process and calibrating it using historical data. I'm not quite sure which model to use for bitcoin. I've seen some papers use a GARCH model. 
In general, am I approaching this correctly? Any suggestions?
Thanks! 
Difficult without understanding the mechanics of the trade that you are suggesting. Generally Vanilla crypto transactions will remove herstatt risk through the distributed GL. There is I suppose the potential that until the currency block as been included up to Level 3 it is not treated as truly completed, so there is a risk that the ledger will reject the transaction.
 
User avatar
bloodynri
Topic Author
Posts: 5
Joined: June 15th, 2011, 12:27 pm

Re: Pricing default risk in cryptos

August 16th, 2018, 4:19 pm

It's actually more like a CDS contract. These are physically settled OTC forwards. It requires initial margin to be put up but there is always a chance a counter-party will default. I want to get insurance against that default and want to know a fair way of pricing it.
 
User avatar
rmax
Posts: 374
Joined: December 8th, 2005, 9:31 am

Re: Pricing default risk in cryptos

August 16th, 2018, 4:49 pm

It's actually more like a CDS contract. These are physically settled OTC forwards. It requires initial margin to be put up but there is always a chance a counter-party will default. I want to get insurance against that default and want to know a fair way of pricing it.
Why not exchange Variation Margin?
 
User avatar
bloodynri
Topic Author
Posts: 5
Joined: June 15th, 2011, 12:27 pm

Re: Pricing default risk in cryptos

August 16th, 2018, 6:22 pm

There will be variation margin as well. Just accounting for the tail risk.
 
User avatar
rmax
Posts: 374
Joined: December 8th, 2005, 9:31 am

Re: Pricing default risk in cryptos

August 16th, 2018, 6:28 pm

There will be variation margin as well. Just accounting for the tail risk.
OK so the Variation Margin will cover you for the default, and will enable you to buy the replacement trade. The Initial Margin accounts for the risk that should the market move hugely when the counterparty defaults then the VM will not be enough. This has nothing to do with the counterparty's credit rating and is much more closley aligned with VaR as it is a market risk. In essence determine a methodology to calculate the VaR of the trade over the horizon that you think it will take to settle the dispute (the SIMM uses 10 days, but possibly in Crypto you may want longer).
 
User avatar
bloodynri
Topic Author
Posts: 5
Joined: June 15th, 2011, 12:27 pm

Re: Pricing default risk in cryptos

August 16th, 2018, 7:02 pm

Agreed. I wanted suggestions to calculate the VaR. I was thinking of modeling bitcoin using some stochastic process (not sure which) and using monte carlo to get a measure of VaR?
 
User avatar
Alan
Posts: 2958
Joined: December 19th, 2001, 4:01 am
Location: California
Contact:

Re: Pricing default risk in cryptos

August 16th, 2018, 7:27 pm

Just googling, I see there is at least one bitcoin options market: here

Given that, one can estimate the associated risk-neutral probability density for each option expiration and calculate tail probs directly. The method I like is to fit IV(K) with a smooth, arb-free form (say J. Gatheral's SVI) and then get the implied density from Black-Scholes + chain rule for [$]C_{KK}[$]. It's more or less a model-free and static method so much preferred to trying to postulate a stochastic process. Essentially, it's an implementation of the Breeden-Litzenberg formula. 

Also, casual observation suggests the risk in bitcoin is largely idiosyncratic, so maybe P-probs = Q-probs is not a bad assumption. 

If you're unfamiliar, there are examples (using SPX and VIX) on pgs 160-162 in "Option Valuation under Stochastic Volatility II".