This post will describe the Portfolio Margin system used at Deribit, a bitcoin derivatives exchange. Portfolio margin is a risk-based margin policy that aligns margin requirements with the overall risk of a portfolio. Portfolio margin usually results in significantly lower margin requirements compared with the sum of margin requirements on individual positions. If a portfolio is long and short, portfolio margin can allow the risk of positions to be “netted off” which may result in an lower overall margin requirement.
I’m most familiar with the CME SPAN margin system which is used to evaluate the risk of futures and options portfolios, but the same general principals apply to Deribit portfolio margin.
Deribit users can request to have portfolio margin applied to their account by e-mailing firstname.lastname@example.org. To qualify, Deribit users must maintain a minimum net equity of at least 0.5 bitcoins, must demonstrate some experience trading options, and declare to have understanding about the concept of portfolio margin.
Here is how Deribit determines portfolio margin:
- maximum price move of +/- 10.00%
- maximum implied volatility change of SQRT (30/days to expiration)*27.00%. Example: options expiring in 30 days: IV change of maximum 27.00%, options expiring in 15 days: IV change of maximum SQRT (30/15)*27.00%
- Contingency component of 0.5% of underlying value of all options in portfolio. Example: you have 200 options in your position (long and short), 0.5% of 200 BTC = 1 BTC is added to the portfolio margin calculation.
- Contingency component of 1.00% of underlying value is added for offsetting futures. Example: you are long 100 BTC in Future A, and short 100 BTC in Future B, then 1.00%*100 BTC will be added to the portfolio margin calculation.
- Contingency component of 0.00% for VEGA’s offsetting in different expirations. Example: you are net long 10 VEGA in expirations A/B/C, and net short 10 VEGA in expirations D/E/F, we will add a contingency of 0.00% thereof to the portfolio margin calculation.
- Initial margin is Maintenance Margin + 30%. Example: If Maintenance Margin is 10 BTC, Initial margin will be 10 BTC+30% = 13 BTC.
Deribit also notes that when liquidating positions, it will start with futures first. This is presumably because the futures contracts are more liquid. Deribit also says that when margin levels are breached, they may liquidate futures first, but may also add futures positions if it reduces overall portfolio risk. For example, if a portfolio is net short puts, Deribit may add short futures instead of buying back puts to achieve minimum margin requirements.
The value of bitcoin plunged 20% against the value of USD overnight, and some of the typical outcomes are happening. The rate to borrow/lend USD on Bitfinex is currently around 87%. This means someone holding $100,000 on Bitfinex is earning $230 per day of interest. Keep in mind Bitfinex uses USD tether, and the best way to withdraw it is to convert to bitcoins and withdraw those, but 87% is still a very high number.
I’m surprised the rate to borrow/lend bitcoins for margin on Poloniex is still very low. The rate has barely changed as the price of bitcoins has gone up and now down over the past few weeks. I would have expected there to be lot’s more margin demand as the price of bitcoin has been increasingly volatile.
News of volume increasing on options exchanges should also be good for traders. On Deribit, implied volatilities spiked overnight. The implied vol on offers is well over 200% today, this is an opportunity for bitcoin holders who want to trade some time/volatility today in exchange for some limited upside in the future.
The order book on QuadrigaCX is full, and shows similar liquidity to the past few weeks. The price of bitcoin on QuadrigaCX is actually higher than the price on rival exchanges. At the time of writing, there is an opportunity for traders to earn arbitrage profits by selling bitcoin on QuadrigaCX for $12,700 and buying bitcoin on Poloniex at $12,562 and/or Bitfinex at $,12,540. The spread is about 1.5% which covers the fees to make the trade. This should be a golden opportunity for those of us running market making bots.
Deribit is a crypto-derivatives site featuring bitcoin futures and bitcoin options. Its a pretty cool site that offers some unique features unavailable to retail users in fiat economies. From a retail user’s perspective, being able to enter limit orders based on either the bitcoin value or the implied volatility value is a cool boost. When users enter an order based on an implied volatility number, the exchange automatically refreshes the order each 6 seconds based on the current variables. This way, a retail trader can enter an order that adjusts to the current market based on a fixed implied volatility number. This is something that most retail brokers in fiat economies do not offer. This type of functionality is obviously available to anyone accessing fiat exchanges using APIs: they simply need to write these crons into their own programs.
I’ve been a fan of trading tail risks ever since the days of InTrade when I wrote the tail risks on economic numbers each night for over a year and never had a losing trade. Its well documented today, with the popularity of behavioural economics over the past few decades, that the untrained human brain makes inaccurate estimates of long shots, and a small mis-estimate for a long shot can translate into a lot of missing/added value when the statistics don’t support such prices. For example, casinos are able to get a bigger house edge for long shot bets compared to even money bets. Consider the high house edge for most land based keno — which is pure long shots — compared to the low house edge for baccarat, the core game of which is close to even money.
Let’s apply the idea that we don’t estimate tail risks accurately to look for ways to profit on Deribit.
Deribit lists serial expirations on monthly and weekly bitcoin options. As time goes by, and expirations get closer, many of the “deep out of the money” strike prices lose liquidity because the minimum tick value is greater than the theoretical value of these tail risks. These deep out of the money strikes will go “no bid” once their theoretical value is less than the minimum tick. This is where the opportunity to profit can be found. I’ve noticed that what I’m assuming are manual order entry retail users with bids posted in the deep out of the money strikes where the theoretical price is lower than the single tick value. I assume that these users entered limit orders without expiration dates and didn’t cancel those orders as the particular strike went “no bid”. Another explanation is a user is short a particular deep out of the money strike and rather than waiting till expiration, the user is willing to pay an above market rate to close out the worthless position in order to free up margin or clean their position book. Whatever the reason, these bids are “pennies from heaven” for the savvy trader.
See the screen shot below that shows the bid on a worthless option at 0.0002 btc. The theoretical value of this option is actually worth less than the minimum tick of 0.0001 btc. A word of warning for those thinking of playing the tail risks: make them covered and remember the old saying about picking up pennies in front of a slow moving steam roller. Most of the time you’ll be able to take the penny, but there is a long-shot chance that you get called away or put in, so you need to either have the capital or enough liquidity to cover (either buying back or hedging). The nice thing about Deribit options is they are settled based on a futures contract that is liquid on the same exchange.
Deribit recently announced they have suffered losses due to liquidations triggered by margin calls. Essentially what happened was as the market prices changed, margin calls and forced liquidations were triggered, and the market did not have enough liquidity to fill all the orders in an orderly manner, so as positions were blown out, there was a big gap in fill prices to where theoretical prices might be. The beneficiaries were the market makers who let the market gap down, then filled margin liquidation orders at prices well below expected prices. The exchange made up the difference (since the liquidated positions were the result of margin calls), and then asked major market makers to eat some of the loss, which it sounds like they did. Below is a copy of the text provided by Deribit.
The case highlights another situation where market participants are at risk since crypto financial intermediaries such as exchanges provide users with too little information about their financial position. With the absence of clearinghouses or independent rating agencies, users are left bearing a lot of risk, and its difficult for users to guage the magnitude of this risk.
Yesterday around 14.00 UTC we had liquidation algorithms of portfolio margin users creation a chaos in the 29 December future. This resulted finally in bankruptcies of more than 105 BTC. Further various client accounts had unjust losses due to liquidations as well.
We had to halt trading yesterday for a while to fix the issue before we could continue again. We are sorry for the downtime.
We solved the issue of the losses by contacting our biggest market makers and traders that have been making profits trading against the malicious algorithm at prices far above the market. We are grateful for their understanding of the incident and for their direct support of our exchange.
Further we decided to refill the insurance fund further such that all other traders will remain completely unaffected and no profits will be socialized among other traders in this session at all.
The total final loss left for the exchange amounts to around 60 BTC (or USD 235.000 at the time of writing). Please note that all users’ funds are safe and we as an exchange can, of course, handle a loss of 60 BTC. The exchange will continue operating as normal.
This was our first major incident since we opened doors for trading in the summer of 2016. We will work hard now to improve various liquidation algorithms such that this could never occur again. This might further delay the launch of new products like Ethereum futures and our upcoming Spot Exchange.
The insurance fund will also be replenished again with 25 BTC.