4370, 1st Semester 2015 - 2016 C U H K Fundamentals of Derivatives Trading Applying Options Theory in the Marketplace Tobias Hekster -Co-CIO and Senior Strategist, True Partner Fund -Adjunct Professor, National Taiwan University 1 © True Partner Education Ltd 2014 C U H K 4370: Fundamentals of Derivatives Trading Strategies Lecture 1 Introduction to the Course Introduction to Trading Markets in Practice Logistics of Markets (market microstructure) Futures and Futures strategies 2 © True Partner Education Ltd 2014 C U H K Fundamentals of Derivatives Trading Strategies Course Introduction • Introduction of the Lecturers • Course Structure and Objectives • Group Projects 3 © True Partner Education Ltd 2014 C U H K Fundamentals of Derivatives Trading Strategies Introduction of the Lecturers: Tobias Hekster Personal Activities past 16 years: 4 1996: Master’s Degree in Economics, Rijksuniversiteit Groningen, The Netherlands 1997: Finance Controller, Royal Dutch Shell, Rotterdam, The Netherlands 1998: Options Trader, IMC Trading, Amsterdam, The Netherlands 2000: Established Cash Arbitrage Desk, IMC Trading 2003: Head-of-Trading, Holland Trading House, Chicago, USA 2008: Head of Volatility Arbitrage, IMC Asia Pacific 2011: Co-CIO and Senior Strategist, True Partner Fund Adjunct Professor Chinese University of Hong Kong Adjunct Professor National Taiwan University © True Partner Education Ltd 2014 C U H K Fundamentals of Derivatives Trading Strategies The aim of the course • Understanding of both the theoretical foundation of Derivatives Trading as implementing trading strategies in practice • Lessons will be consist of a theoretical part where we will discuss Options Theory and trading strategies • The theory will be applied in practice, using various actual trading tools, combined with real-time market data • There will be an examination at the end of the course, which accounts for 65% of the total score • The remaining 35% will be based on group trading projects (30%) as well as class participation in general (5%) 5 © True Partner Education Ltd 2014 Fundamentals of Derivatives Trading Strategies C U H K Group Projects Group Project I: creating a structured product • You will create a derivative structure on an underlying instrument of your choice and write a ‘termsheet’ for retail investors. •Price the option components and calculate profit profile for the customer (and of course for the issuer) 6 © True Partner Education Ltd 2014 Fundamentals of Derivatives Trading Strategies C U H K Group and Individual Projects Group Project II: Volatility Arbitrage Strategy • select strategy and underlying product(s) • calculate expected and realized results • deliverable to be a ‘project proposal’ for the Head of Trading 7 © True Partner Education Ltd 2014 The evolution of Trading C U H K © True Partner Education Ltd 2014 The evolution of Trading C U H K © True Partner Education Ltd 2014 C U H K Introduction to Markets • 10 What are we looking at? © True Partner Education Ltd 2014 C U H K What is a Market? A market is nothing more than a gathering of different opinions on the value of the traded subject. A PRICE is a valuation attributed to an asset by any market participant, THE PRICE is a valuation of the asset that all participants agree with. Central is the process of Price Discovery: If market participants disagree on valuation, they will exchange the traded object. The market clears if no participants are willing to trade with each other anymore. The actual process how buyers and sellers are matched can differ per market and per instrument traded. This is described as market (micro)structure and is an important aspect of today’s trading 11 © True Partner Education Ltd 2014 C U H K Price Discovery in Practice Make an estimate for the number of McDonalds ‘restaurants’ in Hong Kong. The different phases of the auction will show different processes of price discovery 1st Phase: Open Auction Each participant is able to submit a bid for the number of McDonalds’s. - The highest bidder becomes the first buyer - The ‘stand’ price for lower bidders remains in the book This is the standard ‘winner takes all’ variety, most commonly known from the world of art (or overpriced land for real estate development) 12 © True Partner Education Ltd 2014 C U H K Open auction Trader 13 Bid Price Trader Bid Price © True Partner Education Ltd 2014 C U H K Price Discovery in Practice • 2nd Phase: Book building process The book on the buy side consists of the remaining ‘stand’ prices - Bidders can join on the auction price - Sellers can indicate they think the auction price is too high If there is a surplus of sellers at the auction price, we look at the next ‘stand’ price and repeat the process. Auction matches: - When there is an exact balance of buy and sell interests - At the price where the most buys and sells are matched This process is applied at most markets in the ‘pre-opening auction’. Instead of direct matching, there is more room for price discovery prior to matching trades (for example to digest overnight events and movement abroad) 14 © True Partner Education Ltd 2014 C U H K Book Building Trader 15 Bid Price Trader Ask Price © True Partner Education Ltd 2014 C U H K Price Discovery in Practice 3rd Phase: Continuous Trading After the auction match, the book consists of: - The bids that were below the auction match - The offers that were above the auction match All participants are allowed to change their mind and unwind their previous trade. Participants who have not traded yet can decide to buy or sell on this market. At this point, you are allowed to activate your smartphone !!! This trading phase mostly lasts for the remainder of the trading day. It is noticeable how new information changes the market: changes in valuation are immediately reflected and processed in the market 16 © True Partner Education Ltd 2014 C U H K Price Discovery in Practice Let’s replace the number of McDonalds’ with the value of one share in a company. As opposed to a verifiable number (www.wikipedia.com) the value of a share represents the stream of future cash flows, which are uncertain. Thus each one’s valuation is an expectation and an opinion of the prospects of the company. If your valuation exceeds the best offer you should buy shares, if your valuation is below the best bid you should sell shares 17 © True Partner Education Ltd 2014 C U H K Price Discovery in Practice While physically things have changed (electronic), the process does not differ from the old town square markets of the middle ages, there are differences between markets What makes a ‘good’ market: Number of orders in the market (more different opinions, for better price discovery) Spread between the best bid and best ask price (it takes a smaller difference of opinion to trade) Size of orders in the market (availability of volume to trade) Depth of the market (if I want to trade for size, how much more do I need to pay) Transparency (are there hidden traders out there) 18 © True Partner Education Ltd 2014 C U H K Actual Order Books A ‘liquid’ stock: Royal Dutch Shell 19 © True Partner Education Ltd 2014 C U H K Actual Order Books (Cont.) A less liquid stock (the not so famous Dutch building firm Heijmans NV) 20 © True Partner Education Ltd 2014 C U H K Order Entry and ‘lingo’ Suppose we disagree with the displayed market and think the price of the share should be higher. (let’s look at the Heijmans example) What to do: Enter a bid to buy 1 share, whatever the price - ‘market order’ Enter a bid to buy 1 share at EUR 15.635 - ‘lifting the offer’ Enter a bid to buy 1 share at EUR 15.535 - ‘diming the bid’ Enter a bid to buy 1 share at EUR 15.53 - ‘joining the bid’ Enter a bid to buy 1 share at EUR 15.51 - ‘behind the market bid’ Each of these actions carries a different risk-reward pattern: The higher the bid price, the larger the certainty of execution (but at a cost) The lower the bid price, the more favorable the trading price will be (but only if executed) This is called the cost of immediacy and is a key factor in all trading decisions 21 © True Partner Education Ltd 2014 C U H K Crossing the bid-ask: the cost of immediacy Optimal trading strategies: factors that influence trading choice Expected cost Higher Risk Aversion Decreasing order size Shorter trade horizon Increasing Liquidity Increasing Volatility Lower Risk Aversion Increasing order size Longer trade horizon Decreasing Liquidity Decreasing Volatility Timing risk 22 © True Partner Education Ltd 2014 C U H K Crossing the bid-ask: the cost of immediacy Optimal trading strategies: factors that influence trading choice 1) Risk Aversion: to what extent am I prepare to have uncertainty over my order execution 2) Order Size: how many shares do I need to trade (for example, I need 5,000 shares of Heijmans) 3) Trade Horizon: how long do I have to execute my order 4) Liquidity: how much does historically trade in this stock. 5) Volatility: how much does this stock bounce around 23 © True Partner Education Ltd 2014 C U H K Structure of a Trading Day Opening Auction • Overnight events drive greater need for Price Discovery Order accumulation period Continuous dissemination of indicated opening price (or the entire book) 1 matching moment, either random or fixed Continuous Trading • Order book and Trade history suffices for Price Discovery Immediate matching of trades ‘Time/price priority’ or ‘Allocation based’ Circuit breakers • The system should protect against deviations caused by sudden and large imbalances in supply and demand Static Bandwidth Dynamic Bandwidth Volatility Interruptions 24 © True Partner Education Ltd 2014 C U H K Introduction to Markets • Why are markets important? Aggregation of opinions results in optimal results (i) Put your money where your mouth is (ii) Efficient allocation of objects to the highest utility (iii) More participants allow for potential counterparties Sounds straightforward, but… 25 © True Partner Education Ltd 2014 Introduction to Markets C U H K • Put your money where your mouth is! Where does this stock trade? Firm Name CA Cheuvreux AlphaValue ING Wholesale Banking Oddo & Cie Macquarie Exane BNP Paribas Standard & Poor's Equity Research RBS Deutsche Bank Petercam Rabobank International Societe Generale EVA Dimensions Mediobanca SpA Barclays Capital Keefe, Bruyette & Woods Kepler Capital Markets ABN Amro Bank N.V. Natixis Bank Degroof(ESN) Nomura Citi JPMorgan Goldman Sachs 26 Analyst PLUIJGERS NIJDAM PLOEGH SASSUS STOEGNER BOISSIN SILVERMAN NAGTEGAAL BENHAKOUN DE WIT KLUIS LEMAIRE TEAM COVERAGE ROVERE SIGEE LAMBERT PETRARQUE WEIDEMA KOAGNE LEEMANS JOSHI NEDIALKOV FORMANKO NEUEZ Recommendation underperform buy hold neutral neutral neutral hold hold hold add buy hold overweight outperform overwt/neutral underperform buy buy neutral hold buy buy overweight not rated M M M M M M M M M U M M M M M M M M M M M M M M Tgt Px 10 17 12.5 29.5 15 15 17 12.5 15.8 14.2 25 20 18 15 12.6 32 22 15.5 20 36 40 40 © True Partner Education Ltd 2014 C U H K Introduction to Markets Put your money where your mouth is - US elections: opinion polls vs online-betting - Credit Ratings (remember AAA rated subprime CDO’s) There is a difference between the value one disseminates, and the price one is actually prepared to buy or sell an object at. As opposed to talking, in trading the validity of one’s value and thus the traded prices define profit. This course will define trading relating to the perspective of Market Making, Liquidity Provision (which differs from the end-user) 27 © True Partner Education Ltd 2014 C U H K The basics of Trading As soon as one has a valuation of an object, one can trade • Taking risk for a premium (or credit, buying the object below the theoretical value or selling it above) • Trying to dispose of that risk and lock in (part of) the premium Credit: difference between our valuation and the current market price Risk: possibility and magnitude of unfavorable outcomes © True Partner Education Ltd 2014 C U H K The basics of Trading: Risk The following formula roughly applies in general for measuring risk (using normal distributions): Risk = Value of Position * Volatility * SQRT (Time) Please note: • The longer the holding period, the increase of the risk is not linear • Watch out for definitions of risk: forward looking standard deviation is reflective, but backward looking it is not • This model is applied by most banks (Value at Risk) and has some flaws => The willingness to take risk is Risk Tolerance © True Partner Education Ltd 2014 The basics of Trading: Credit C U H K • Premium = ‘credit’; difference between your valuation and the current market price • A market price that is out of line according to your pricing model • No premium => no trade, only for gamblers • Yes! premium => trade if Premium (‐ Cost) >= Risk Aversion * Risk © True Partner Education Ltd 2014 The basics of Trading C U H K “Am I so clever or are they so stupid?” • Different traders have different goals. It's not smart people versus stupid people in the marketplace. • You can have two traders each at the opposite side of a trade and both make money because the scope of the trade may be different. • Beware of the ‘greater fool’ theory • There is no such thing as a free lunch (it’s all about risk / reward) © True Partner Education Ltd 2014 The basics of Trading: A rigged coin flip C U H K • It is all about managing your risk and position • Even with trading with credit al the time credit you can still go bankrupt! • As an example: a rigged coin flip game: – 60% chance of heads – 40% chance of tails © True Partner Education Ltd 2014 Trading Strategies: a rigged coin-flip gain C U H K Assume you can participate in the following game: • Flip a coin that has a 60% to return Tails • You can play the game an endless amount of times • But you have to invest a fixed proportion of your capital each flip • What percentage should I invest (or rather, bet)? 33 © True Partner Education Ltd 2014 Trading Strategies: a rigged coin-flip gain C U H K So even the best trade can kill you, when not properly managed… (by the way, the answer is about 20% according to the Kelly Criterion) 34 © True Partner Education Ltd 2014 C U H K Market Micro Structure • Market structure and design • Trading Mechanism • Fees and Transaction Costs © True Partner Education Ltd 2014 Markets Micro Structure C U H K Market Type • Quote Driven • • • • Continuous display of bid- and ask prices Take-it or leave it / Limited negotiation Preference / Benefit of Quote provider Order driven • • • All combined orders create the order book Lack of orders hinders price discovery All orders created equal (pure time-price or allocation) © True Partner Education Ltd 2014 Markets Micro Structure C U H K Order Types (the more exotic ones) • • • • • • • Immediate or cancel Fill or kill Icebergs Pegged orders Flash orders Routing orders (fragmented markets) Intermarket Sweep orders (fragmented markets) © True Partner Education Ltd 2014 Order Types C U H K Assume the orderbook is as follows: Bid Size Bid Price Ask Price Ask Size 500 251 252 300 100 250 253 600 1,000 249 254 300 The difference between IOC and FOK is as follows: Buy 500 at 252 as type IOC will buy 300 and the remainder will be cancelled Buy 500 at 252 as type FOK will not execute and be cancelled © True Partner Education Ltd 2014 Order Types C U H K Assume the orderbook and trade list are as follows: Bid Size Bid Price Ask Price Ask Size Time Quantity Price 1,500 251 252 300 01:24:10 300 252 1,000 250 253 600 01:24:50 600 252 1,000 249 254 300 01:25:30 300 252 01:26:10 600 252 How to spot an Iceberg: - Quite a lot seems to be trading at 252 in multiples of 300 - Does the 300 offer at 252 remain even after it traded - If the offer at 252 would be for a far larger size, does that impact your decision if you are offering at 253 => Analysis of the order book is vital to trading (and automatable) © True Partner Education Ltd 2014 Order Types C U H K Assume the orderbook is as follows: Bid Size Bid Price Ask Price Ask Size 500 251 252 300 200 250 253 600 1,000 249 254 300 A Pegged order intends to have a bid (ask) price that is related to the best bid (ask) price in the book: In this example, our order will always be: - Our bid = Best Bid – 1 point - Our ask = Best Ask + 1 point Why do this: catch orders that are ‘walking the book’ © True Partner Education Ltd 2014 Order Types: fragmented markets C U H K Assume the orderbooks are as follows: Market ABC Market XYZ Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.50 300 1,000 250.00 251.00 1,000 200 250.00 252.00 600 700 249.50 251.50 500 1,000 249.50 252.50 300 300 249.00 252.00 400 When I send a buy order to market ABC, I could actually get better execution at market XYZ. So I need to make a decision: - What market to route my order to? - Do I want to expose the order first in ABC and if no interest route to XYZ? - If my order is larger than 1,000 shares, do I split the order? © True Partner Education Ltd 2014 Order Types: fragmented markets Market ABC C U H K Market XYZ Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.50 300 1,000 250.00 251.00 1,000 200 250.00 252.00 600 700 249.50 251.50 500 1,000 249.50 252.50 300 300 249.00 252.00 400 - If I expose the order briefly at market ABC, I cannot use an IOC order (as that will immediately cancel) - The concept of a ‘flash order’ was to briefly expose the order to a limited number of participants on market ABC before routing away to XYZ - This caused concerns of ‘frontrunning’ and has been discontinued © True Partner Education Ltd 2014 Order Types: fragmented markets Market ABC C U H K Market XYZ Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.50 300 1,000 250.00 251.50 1,000 200 250.00 252.00 600 2,000 249.50 252.00 2,500 1,000 249.50 252.50 300 300 249.00 252.50 1,400 - If I want to enter a large buy order at 251.00, how should I distribute over the different exchanges? • Based on current markets (XYZ shows more volume) • Based on past statistics, either market could have more volume • How to distribute between markets (‘one cancels the other’) © True Partner Education Ltd 2014 Order Types: fragmented markets Market ABC C U H K Market XYZ Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 5,000 250.50 251.50 5,000 18,000 250.50 251.00 50 200 250.00 252.00 600 2,000 250.00 251.50 500 1,000 249.50 252.50 300 3,600 249.50 252.00 1,400 - If I want to enter a large buy order (6,000) at 251.50: • If I buy the 251.50 offer on market ABC, I have traded through the 251.00 offer on market XYZ (not allowed) • If I buy the 251.00 offer first, I might scare away the offer at 251.50 => The solution is an Intermarket Sweep order, which does exactly that. © True Partner Education Ltd 2014 Order Types: fragmented markets Market ABC C U H K Market KLM (dark pool) Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 5,000 250.50 251.50 5,000 18,000 Mid Mkt Offer 24,000 200 250.00 252.00 600 1,000 249.50 252.50 300 - My buy order in the dark pool is just waiting for any selling interest. As soon as a seller emerges at Mid Market, we would be matched at the Mid Market of the primary market ABC - If someone is interested in buying the offer at market ABC, they should also check the dark pool, as an additional sell order is here at the offer (look at the similarity with a pegged order) © True Partner Education Ltd 2014 Order Types: in-house matching Market ABC C U H K Client Orders Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price 500 250.50 251.50 15,000 100 251.50 200 250.00 252.00 600 1,000 249.50 252.50 300 Ask Price Ask Size - The bank matches my order at 251.50 with their own trading desk - I might get charged lower transaction fees, but the bank has a valuable proposition, as they can trade at the bid and offer in the market, but take the first pick - Complaints that this drains liquidity from the primary market ABC are valid, as well are concerns of best execution and paper trail => In house matching now only performed at price improvement as opposed to just matching the market (SEC proposal) © True Partner Education Ltd 2014 Fragmented markets C U H K Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.50 5,000 1,500 250.50 251.50 1,000 200 250.00 252.00 600 200 250.00 252.00 600 1,000 249.50 252.50 300 1,000 249.50 252.50 300 Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.00 1,000 500 250.50 251.50 600 200 250.00 252.00 600 200 250.00 252.00 600 1,000 249.50 252.50 300 1,000 249.50 252.50 300 Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 15,000 Mid Mkt Offer 5,000 500 251.00 251.50 3,000 4,000 Bid 200 250.00 252.00 600 1,000 249.50 252.50 300 Comparing and routing becomes humanly impossible… The US has now over 50 different venues Bid Size Bid Price Ask Price Ask Size Bid Size Bid Price Ask Price Ask Size 500 250.50 251.50 900 500 250.50 251.50 700 200 250.00 252.00 600 200 250.00 252.00 600 1,000 249.50 252.50 300 1,000 249.50 252.50 300 © True Partner Education Ltd 2014 Markets Micro Structure C U H K Market Transparency • Lit Markets • • • • • Entire order book is displayed Trades follow visible matching of trade intensions Price discovery takes place Primarily limit orders Dark pools • • • • Part or none of the order book is displayed Trades occur if invisible intensions match Depends on price discovery elsewhere Primarily pegged orders © True Partner Education Ltd 2014 Markets Micro Structure C U H K Trading Mechanism • • • Continuous trading (oddly named Continuous Auction) Periodic auctions Request for quotation © True Partner Education Ltd 2014 Continuous Trading - C U H K Orders can be entered into the order book As soon as an entered order equals or crosses an opposite order, the trade is matched Residual orders remain in the order book Bid Size Bid Price Ask Price Ask Size 500 251 252 300 100 250 253 600 1,000 249 254 300 What happens if I enter an offer for 600 shares at 251? What happens if I enter a bid for 500 shares at 250? © True Partner Education Ltd 2014 Continuous Trading C U H K What happens if I enter an offer for 600 shares at 251? What happens if I enter a bid for 500 shares at 250? Bid Size Bid Price Ask Price Ask Size 600 250 251 100 1,000 249 252 300 253 600 What happens if now an offer of 100 is entered at 250? - Time price priority (‘first come, first serve’) - Allocation algorithm (‘all is shared’) © True Partner Education Ltd 2014 Continuous Trading C U H K What happens if now an offer of 100 is entered at 250? Bid Size Bid Price Ask Price Ask Size 100 (bid 1) 250 251 100 500 (bid 2) 250 252 300 1,000 249 253 600 - Time price priority (‘first come, first serve’) Bid 1 executes all 100 - Pro-rata allocation algorithm (‘all is shared’) Bid 1 executes 16, bid 2 executes 84 Due to these incentives, time/price markets tend to be tight for smaller displayed volumes, whereas pro-rata markets tend to be wider for large displayed volumes © True Partner Education Ltd 2014 Auction - C U H K Orders can be entered into the order book As soon as an entered order equals or crosses an opposite order, an indicative matching is disseminated Matching orders remain in the order book At one moment (the Auction Match), matching orders are matched Unmatched orders remain in the order book Bid Size Bid Price Ask Price Ask Size 500 251 252 300 100 250 253 600 1,000 249 254 300 What happens if I enter an offer for 600 shares at 251? What happens if I enter a bid for 500 shares at 250? © True Partner Education Ltd 2014 Auction C U H K What happens if I enter an offer for 600 shares at 251? What happens if I enter a bid for 500 shares at 250? Bid Size Bid Price Ask Price Ask Size 500 251 251 600 600 250 252 300 1,000 249 253 600 Indicative Match Price: Indicative Match Quantity: 251 500 What happens if I enter another offer for 400 at 250? © True Partner Education Ltd 2014 Auction C U H K What happens if I enter another offer for 400 at 250? Bid Size Bid Price Ask Price Ask Size 500 251 250 400 600 250 251 600 1,000 249 252 300 If the matching price would be 250, that means that not all the bids above 250 will get execution at the lower price of 250. That cannot be, thus: Indicative Match Price: Indicative Match Quantity: 251 500 Now another offer appears, at 250 for 200 shares © True Partner Education Ltd 2014 Auction C U H K What happens another offer for 200 at 250? (so 600 offered in total) Bid Size Bid Price Ask Price Ask Size 500 251 250 600 600 250 251 600 1,000 249 252 300 If the matching price would be 251, that means that not all the offers at 250 will get execution at 251. Indicative Match Price: Indicative Match Quantity: 250 600 Within the auction, equal orders again are to be allocated time/price or pro-rata © True Partner Education Ltd 2014 Markets Micro Structure C U H K Transaction costs • Direct transaction costs • • • • • Stamp duty Brokerage fee Exchange fee Clearing fee Indirect transaction costs • • Bid / ask spread Market impact © True Partner Education Ltd 2014 C U H K Futures “A standardized contract between two parties to exchange a specified asset for a price agreed today with delivery taking place at a specified date in the future” Futures are the oldest financial instruments, dating back to ancient Greece (the irony!) with descriptions by Aristotle of a futures contract on Olives for pressing into olive oil. (developed by a philosopher turned trader) All types of underlyings (currencies, commodities, shares, energy-transfercapacity, etc) Index Future: the underlying asset of the future is the value of the index Hang Seng Future S & P 500 Future 58 © True Partner Education Ltd 2014 C U H K Futures vs Forwards: Both instruments are agreements to: - Exchange a certain underlying instrument - At an agreed price - At an agreed point in the future The difference is that Futures are listed and cleared through the facilities of an exchange, whereas Forwards are OTC contracts between two parties => Forwards are subject to counterparty risk (Lehman Brothers) 59 © True Partner Education Ltd 2014 C U H K Futures vs Forwards: The process of Novation: A trade between Buyer A and Seller B is split into two different positions: - Buyer A buys from the Clearing House (‘central counterparty’) - Seller B sells to the Clearing House Buyer A Clearing House Seller B After the Matching of the trade, there is no relation between A and B, both only have counterparty risk towards the Clearing House 60 © True Partner Education Ltd 2014 C U H K Futures: marking-to-market As the Clearing House bears all the counterparty risk, it will ask the traders to post a deposit at the clearing bank: Margin On a daily basis, outstanding contracts will be settled to the closing price of the day (‘marking to market’) Based on the mark to market, a payment will accrue between counterparties of a futures trade, reflecting the daily change in the price of the future For example, if the TAIEX future drops 10 points (contract size NT$ 200) this results in a transfer of NT$ 2,000 from the buyer to the Clearing House and on to the seller Buyer Clearing House - NT$ 2,000 61 Last mark: 7,500 New mark:7,490 Seller + NT$ 2,000 © True Partner Education Ltd 2014 Futures: Contract specifications C U H K Hang Seng Future 62 © True Partner Education Ltd 2014 Futures: Contract specifications (Cont.) C U H K S&P 500 Future 63 © True Partner Education Ltd 2014 C U H K How to price a Future: “Cash and Carry Arbitrage” If you are to deliver the value of an index, one year from now, how to hedge? Buy the index right now! You make an investment today equal to the current index You will incur interest costs over this investment You will receive dividends over the holding period Future = Spot + Interest – Dividends It works the same for receiving the value of the index: You will receive an amount today equal to the current index You will receive interest costs over this investment You will pay dividends over the holding period 64 © True Partner Education Ltd 2014 C U H K How to price a Future: (Cont.) Theoretically, the future should always trade at this level! In practice, there are reasons for deviation: If you want to sell shares, you need to borrow them. This requires a payment, thus your net received interest will be lower Not all dividends are already known, so other participants could have other predictions Stamp duty and other costs (in Hong Kong, 10 bbp stamp duty corresponds with 22 points in the Hang Seng!) 65 © True Partner Education Ltd 2014 C U H K How to price a Future: (Cont.) Hang Seng Future: different maturities 66 © True Partner Education Ltd 2014 C U H K How to price a Future: (Cont.) This formula applies for all different maturities: Future (September) = Spot + Interest (now to Sept) – Dividend (now through Sept) Future (December) = Spot + Interest (now to Dec) – Dividend (now through Dec) Note that the Spot value is a common factor, therefore the difference between the futures - The Future Roll Future Roll (Sept, Dec) = Interest (Sept to Dec) – Dividend (Sept through Dec) If you buy the future roll (that is buy December future, sell September future), you will have the following risks: Change in the interest rate between Sept and Dec Change in dividends between Sept and Dec 67 © True Partner Education Ltd 2014 C U H K Future roll: example In February 2008, the French bank Societe Generale had to announce a massive loss of several billion Euros, as one of their traders had accumulated an enormous rogue position Societe Generale is a large component of the French CAC40 index Societe Generale is the only CAC40 component to pay a dividend in early March If you hear the news hitting the tape, what would you do: a) Buy the March – February Future Roll in the CAC40 (+ March, - February) b) Sell the March – February Future Roll in the CAC40 (- March, + February) c) Do nothing 68 © True Partner Education Ltd 2014 C U H K The Future tail Is there no risk whatsoever to movements in the spot? Let’s look at the formula for the future roll: Future Roll (Sept, Dec) = Interest (Sept to Dec) – Dividend (Sept through Dec) The interest is calculated over the spot price So the higher the Spot, the higher the interest amount! The future can move more than the underlying, this is called the future tail. 69 © True Partner Education Ltd 2014 Future tail example: the infamous Volkswagen saga C U H K In the end of October 2008, Porsche announced their stake in competitor Volkswagen was far larger than previously known. The result was the ‘mother of all short squeezes’ with shares of Volkswagen briefly exceeding EUR 1,000 A market making firm traded long term dividend expectations in Volkswagen by spreading the following position: - Long 100,000 shares of Volkswagen - Short 100,000 stock futures in Volkswagen (1 year maturity) Assume the following: - Interest rate 3% per annum - Dividend of EUR 2.00 paid at the end of 2008 - Share price rises to EUR 1,000 from EUR 300 before the announcement When the local head of trading calls that “there appears to be some money missing in the trading sheet”, was that true and if so, how much? 70 © True Partner Education Ltd 2014 Future tail example: the infamous Volkswagen saga C U H K A market making firm traded long term dividend expectations in Volkswagen by spreading the following position: - Long 100,000 shares of Volkswagen - Short 100,000 stock futures in Volkswagen (1 year maturity) Assume the following: - Interest rate 3% per annum - Dividend of EUR 2.00 paid at the end of 2008 - Share price rises to EUR 1,000 from EUR 300 before the announcement Old futures price: New futures price: Loss on futures: Gain on stock: EUR 300 x 1.03 – 2.00 = EUR 307 EUR 1,000 x 1.03 – 2.00 = EUR 1,028 EUR -/- 72,100,000 EUR 70,000,000 Missing from the sheet: -/- EUR 2,100,000 71 © True Partner Education Ltd 2014 C U H K Future trading strategy: examples Contract Specifications SGX Eurex Tick Size 1 point tick 1 point tick Contract Size USD 10 x index pt EUR 10 x index pt They certainly look suitable for spread strategies: • Translate liquidity by quoting SGX at an edge • Quoting both sides for ‘order book benefit’ 72 © True Partner Education Ltd 2014 C U H K Future trading strategy: examples (Cont.) Quoting a dual listing Eurex Market SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 300 2,801 2,802 150 100 2,800 2,804 150 1,400 2,800 2,803 200 200 2,798 2,805 100 200 2,799 2,804 650 150 2,796 2,806 300 We can ‘lean’ on the Eurex market and allow for 1 tick of credit quoting SGX; When traded, immediately hedge on Eurex. Eurex Market 73 SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 300 2,801 2,802 150 200 2,800 2,803 50 1,400 2,800 2,803 200 200 2,798 2,805 100 200 2,799 2,804 650 150 2,796 2,806 300 © True Partner Education Ltd 2014 C U H K Quoting a dual listing The benefit of this strategy is the translation of liquidity from the main product (Eurex) to the alternative listing (SGX). This in turn provides retail investors with a better market and is thus crucial for the success of the alternative listing One can place multiple quotes at all layers of the book (even reverse) As the pricing is so obvious, this strategy is highly competitive Quoting holds significant exposure: When the market in the liquid instrument changes, one needs to immediately adjust prices in the market When the quoter trades, one needs to immediately hedge in the liquid future It is all about being fast (“ultra-low latency”) 74 © True Partner Education Ltd 2014 C U H K What is ‘Latency’ • Latency: microsecond is 0.000001 seconds……blink of an eye 100 – 150 milliseconds • Latency represents the inherent delays in transmitting and processing data (for orders or market data prices) => Minimize distances involved: Co-location => Optimize internal processing => Keep data messages as small as they can 75 © True Partner Education Ltd 2014 C U H K Dual listings on the same underlying: hitting Quoting a dual listing Eurex Market SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 300 2,801 2,802 150 100 2,800 2,804 150 1,400 2,800 2,803 200 200 2,798 2,805 100 200 2,799 2,804 650 150 2,796 2,806 300 We can ‘lean’ on the Eurex market and pre-define a potential order(s), we can shoot into the market at will: Market Volume Side Price Type SGX 50 Sell 2,803 IOC SGX 50 Sell 2,800 IOC • Order 1 waits for an 2,803 bid to appear on SGX • Order 2 waits for an 2,799 offer to appear on Eurex 76 © True Partner Education Ltd 2014 C U H K Dual listings on the same underlying: hitting The benefit of this strategy is to increase the likelihood of execution for orders In the alternative instrument. But contrary to quoting, the price discovery in The alternative instrument does not benefit Multiple potential orders can be generated, based on all layers of both books. But opposite to quoting, if bid/ask of market is taken, hitter does not participate Hitting can be a predatory strategy Injecting a new order into the market is mostly faster than modifying an existing order (the quoter gets ‘picked off’) The hitter has no obligation or exposure, just lays in ambush It is still all about ultra-low latency 77 © True Partner Education Ltd 2014 C U H K Spreading 2: jumping the queue Quoting a dual listing Eurex Market SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 1,400 2,800 2,801 150 100 2,799 2,802 150 200 2,799 2,802 200 200 2,798 2,803 100 350 2,798 2,803 650 150 2,796 2,804 300 We can ‘lean’ on the bid-size on Eurex and match it on SGX: Eurex Market 78 SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 1,400 2,800 2,801 150 100 2,800 2,802 200 200 2,799 2,802 200 100 2,799 2,803 150 350 2,798 2,803 650 200 2,798 2,804 100 © True Partner Education Ltd 2014 C U H K Spreading 2: jumping the queue (Cont.) When buy 14 futures at 2,800 at SGX we can do the following: • Offer 14 futures at 2,801 on SGX for a ‘scalp’ • Offer 10 futures at 2,801 on Eurex for a ‘scalp’ • ‘Scratch’ by selling 10 futures on Eurex at 2,800 if the market bid shrinks Eurex Market SGX Market Q bid P bid P ask Q ask Q bid P bid P ask Q ask 1,400 2,800 2,801 250 200 2,800 2,801 100 200 2,799 2,802 200 200 2,798 2,802 200 350 2,798 2,803 650 150 2,796 2,803 150 The purchase at 2,800 has a positive expected value ! 79 © True Partner Education Ltd 2014 Dual listings on the same underlying: queue jumping C U H K When we investigate the result of the 3 potential outcomes: Offer 2,801 on SGX for a ‘scalp’ - P&L is USD 140 minus transaction costs * = USD 112 Offer 2,801 on Eurex for a ‘scalp’ - P&L is USD 140 minus transaction costs (applying EUR/USD at 1.40) Sell at 2,800 if the market bid shrinks (‘scratch’) - P&L is -/- the transaction costs = -/- USD 28 * Assumes transaction costs are USD 1 per future (and EUR 1 on Eurex) If this trade works out more than 1-in-4, it has a positive net expected value 80 © True Partner Education Ltd 2014 Dual listings on the same underlying: queue jumping (Cont.) C U H K This ‘second generation’ HFT strategy even further benefits the liquidity in the alternative instrument, as prices could even match the liquid instrument. Requires more parameters to be successful: What size on Eurex is deemed sufficiently large How long should one hold the position before ‘scratching’ When parameters are identical, the fastest execution is the difference. When the bid shrinks, it becomes a game of chicken: who dares to stick to his position the longest… It is even more about ultra-low latency 81 © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? The SGX product is ‘quantoed’: the underlying index is calculated in EUR, but the SGX listing trades in USD. Does that matter? Just a headline in Reuters: GLOBAL MARKETS-Stocks sink, dollar rallies on risk aversion Assume we have the following position: 82 Position Instrument Price Delta Currency Long 100 Future Eurex 2,800 1,000 EUR Short 140 Future SGX 2,800 -/- 1,400 USD © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? Suppose the EuroStoxx drops 50 points, with the Euro/Dollar rate unchanged at 1.40 Position Instrument Result (local) Result (USD) Long 100 Future Eurex -/- EUR 50,000 -/- 70,000 Short 140 Future SGX USD 70,000 70,000 Net Result 0 => The position remains fully hedged 83 © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? Suppose the EuroStoxx drops 50 points, but now the previous headline materializes and the Euro/Dollar drops to 1.38 Position Instrument Result (local) Result (USD) Long 100 Future Eurex -/- EUR 50,000 -/- 69,000 Short 140 Future SGX USD 70,000 70,000 Net Result 1,000 => We earn money as the loss is in the depreciating currency and the profit is in the appreciating currency 84 © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? But in the past, what was good for the US Dollar, was good for the European exporters. Suppose in this regime, the EuroStoxx rises 50 points combined with a drop in Euro/Dollar to 1.38 Position Instrument Result (local) Result (USD) Long 100 Future Eurex EUR 50,000 69,000 Short 140 Future SGX -/- USD 70,000 -/- 70,000 Net Result -/- 1,000 => We lose money as the loss is in the appreciating currency and the profit is in the depreciating currency 85 © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? In the below graph, all scenario’s are depicted: the result in Index Points of different movement results 15 10 5 Change EUR/USD 0 -5 3.75% 2.25% 0.75% -0.75% -10 -15 -2.25% -3.75% Change in Stoxx 50 For example, if the index moves up 1.5% and the Euro moves up 0.75% the effect is a gain equivalent to 0.5 index points 86 © True Partner Education Ltd 2014 C U H K Appendix: twist to the previous dual-listing? So these effects, while relatively small on a day-to-day basis, could add up over the life of a future contract. While symmetric, the results do become more pronounced in times of volatility If a correlation (‘risk-on, risk-off’) is present there should be a premium/discount between USD and EUR denominated futures Imagine the earthquake in Japan last March and the effect on dual-listed Nikkei futures (in JPY or USD): a 10% move in the Nikkei and a 4% move in the USD/JPY exchange rate © True Partner Education Ltd 2014 C U H K The Concept of Pairs Trading Every stock has its alpha • The ‘classic’ Capital Asset Pricing Model has beta as ‘just’ the indexrelated performance and alpha as specific (out)performance • While we see this terminology commonly in hedge-fund lingo, the key underlying concept is that instruments have a tendency to move in a correlated fashion • This moving in tandem mostly has obvious reasons - Similar type of companies (Bank of China vs ICBC) - Non-fungible dual listings (Swire ‘A’ shares vs Swire ‘B’ shares) - Company and their products (Exxon Mobile vs West Texas Intermediate (crude)) 88 © True Partner Education Ltd 2014 The Concept of Pairs Trading C U H K How to benefit The ‘classic’ Pairs trade aims to benefit from the fact that instruments move in tandem: When instruments move purely in tandem, they are near perfect hedges • Which one is cheaper to trade (bid/ask spread) • We can provide liquidity in the wider instrument When the individual instruments’ deviation can be assumed to be noise • The relationship between their prices will be mean reverting • We can trade the mean reversion and amplitude embedded in this noise In both cases, our position will be spreading the two instruments General market exposure (‘beta’) is minimized The position will be driven by specific exposures (‘alpha’) 89 © True Partner Education Ltd 2014 The Concept of Pairs Trading C U H K The alpha slide revisited While we see this terminology commonly in hedge-fund lingo, the key underlying concept is that instruments usually have a tendency to move in a correlated fashion This moving in tandem mostly has obvious reasons Similar type of companies (Sun Hung Kai vs Cheung Kong) Non-fungible dual listings (Volkswagen shares vs Volkswagen preferreds) Company and their products (BP vs West Texas Intermediate (crude)) Single stocks are more susceptible to shock events than indices Over the past years, correlations between equity indices were at elevated levels (within the Euro zone economies above 90%) 90 © True Partner Education Ltd 2014 C U H K Spreading two correlated instruments We can apply the same logic to futures on different indices, as long as the relation between them is tight • Consider the two main European index futures: DAX EURO STOXX 50 Constituents 30 German 50 Eurozone Cross Members 14 14 Cross Weight 78 % 34 % • Of the main, liquid, European indices, this pair exhibits the highest correlation (0.956 over the past year) 91 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K Once we have established which instruments to trade, how to build the strategy from here: 1) Which leg shall we quote in, which leg will be our hedge 2) What is the (theoretical) relationship between the two legs 3) Define trading parameters: • Credit • Position adjustment • Time-based adjustment 4) ‘Backtesting’ and Risk Tolerance 5) Practical considerations 92 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K Selecting the Underlying and the ‘Quotee’ If we observe the two contracts: DAX EURO STOXX 50 Contract Size EUR 25 per point EUR 10 per point Tick Size 0.5 point 1.00 point Last Traded 6,763 2,313 Contract Value EUR 169,075 EUR 23,130 Spread (bps) 0.7 bps 4.3 bps In order to benefit from quoting a wide market, we quote in the EURO STOXX and hedge in the DAX For a similar contract value on both legs, we will trade 1 DAX future vs 7 EURO STOXX 93 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Setting a (theoretical) relationship The next step is to define a theoretical level, to base our initial quotes on: Bid Price Ask Price DAX 6,800 6,800.50 Euro Stoxx 50 2,311 2,312 Applying a straight ratio, we could define the following relationship: Euro Stoxx = 0.34 x DAX What does the ratio indicate: Ratio’s above 0.34: Euro Stoxx rises and/or DAX declines since our fixing Ratio’s below 0.34: Euro Stoxx declines and/or DAX rises since our fixing 94 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Setting a (theoretical) relationship Assuming the following markets in the futures: Bid Price Ask Price DAX 6,800 6,800.50 EURO STOXX 50 2,311 2,312 Applying the ratio of 0.34, we would get the following theoretical pricing: Bid Price Ask Price EURO STOXX 50 (actual) 2,311 2,312 EURO STOXX 50 (theoretical) 2,312 2,312.17 => So should we buy the Euro stoxx future at 2,312 (and sell the DAX future at 6,800?) 95 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters On the previous slide, our theoretical bid price matched the offer in the market, but… - The objective is to earn money, which trading at theoretical will not help in - There are transaction costs and other frictions involved - Quoting (and trading in general!) carries risks, which we intend to be rewarded for We should take a ‘credit’ every trade we do; let’s assume 2 index points in the Euro Stoxx futures 96 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Mechanics of a spread trading strategy: theoretical pricing Looking again at the market in the futures: Bid Price Ask Price FDAX 6,800 6,800.50 SX5E 2,311 2,312 Applying the ratio of 0.34, and our 2 points credit, we calculate * SX5E Theoretical Bid Theoretical Ask 2,310 2,314.17 * Calculation: Theo Bid = Ratio x FDAX bid – credit: 0.34 x 6,800 – 2 = 2,312 97 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters Assume the same market in the futures: Bid Price Ask Price DAX 6,800 6,800.50 EURO STOXX 50 2,311 2,312 Applying the ratio of 0.34 as well as a credit of 2 points, we would get the following theoretical pricing Bid Price Ask Price EURO STOXX 50 (actual) 2,311 2,312 EURO STOXX 50 (theoretical) 2,310 2,315* The calculation remains straightforward: Theo Bid = Ratio x DAX Bid – Credit (0.34 x 6,800 – 2 = 2,310) * As the tick size is a full point in the EURO STOXX, we should round-up our theoretical price 98 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters Assuming the same market in the DAX futures, our quote will be as follows: Bid Price Ask Price FDAX 6,800 (Market) 6,800.50 (Market) SX5E 2,310 (Our) 2,315 (Our) In our dreams, assuming an unchanged market in the DAX future, we would trade on both sides of our quotes (‘the noise happens in the Euro Stoxx future’) 99 Buy 7 SX5E 2,310 Sell 1 FDAX 6,800 Sell 7 SX5E 2,315 Buy 1 FDAX 6,800.50 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters In this fine scenario, we make one full ‘scalp’ In the Euro Stoxx, we scalp 5 points over 7 futures: 5 x 7 x EUR 10 = EUR 350 In the DAX, we have to give away the bid/ask spread over 1 future: 0.5 x 1 x EUR 25 = -/- EUR 12.50 overall, one full scalp earns us EUR 337.50 In terms of the ratio, we do two trades: • Buy the ratio at 0.3397 (which is 2,310 / 6,800) • Sell the ratio at 0.3404 (which is 2,315 / 6,800.50) 100 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters As in the previous slide, we start with buying the ratio... Buy 7 EURO STOXX 50 2,310 Sell 1 DAX 6,800 But after these two trades, the market remains as is: Bid Price Ask Price DAX (market) 6,800 6,800.50 EURO STOXX 50 (us) 2,310 2,315 EURO STOXX 50 (market) 2,309 2,310 So what is next? 101 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters We now have a position in the spread (as we did a ‘half scalp’) The goal of our strategy is to capture the bid/ask spread and trade the ‘noise’ in the spread. Therefore, we do need to continue quoting in the market. As the bid in the DAX is still 6,800, our theoretical bid would still be 2,310, but with this bid we will trade another half scalp, do we want that? With the offer in DAX at 6,800.50, our theoretical offer would still be 2,315 but would we be prepared to accept a lower price, to unwind our risk? If the answer to both questions is “yes”, we have intuitively changed our valuation of our position. 102 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters As this strategy is not a pure arbitrage (the 0.34 ratio is not set in stone), we do not have infinite risk appetite. •We are only prepared to accumulate more position at a better price •We are prepared to unwind our position at a lower credit The position should be reflected in our theoretical pricing: “for every 7 Stoxx futures in my portfolio, I reduce my theoretical value with two index points” This concept is the cornerstone of continuous trading strategies and referred to as Inventory Based Pricing 103 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters Applying Inventory Based Pricing results in the following new quote: Bid Price Ask Price Position DAX (market) 6,800 6,800.50 -1 EURO STOXX 50 (market) 2,309 2,310 7 EURO STOXX 50 (theo) 2,308 2,313 Our theoretical bid and ask price are calculated as follows: Theo Bid = 0.34 x Bid (DAX) – Credit – 2x Position (SX5E) / 7 Theo Bid = 0.34 x 6,800 – 2 – 2 x 7/7 = 2,308 Theo Ask = 0.34 x 6,800.50 + 2 – 2 x 7/7 = 2,313 (rounded up) 104 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters If we would make a full scalp afterwards, the revenue has changed: Buy 7 EURO STOXX 50 2,310 Sell 1 DAX 6,800 Sell 7 EURO STOXX 50 2,313 Buy 1 DAX 6,800.50 We ‘scalp’ 3 points in the 7 Euro Stoxx futures (EUR 210), but we incur a cost of 0.5 points in the DAX future (-/- EUR 12.50) for an overall result of EUR 197.50 105 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters While the Inventory Based Pricing avoids immediate accumulation, it does not eliminate position risk: Suppose the Euro Stoxx slides lower, while DAX remains unchanged… (Spain anyone?) 106 Trade Price Position Av. Price Profit 7 2,310 7 2,310 0 7 2,308 14 2,309 -/- 140 7 2,306 21 2,308 -/- 420 7 2,304 28 2,307 -/- 840 7 2,302 35 2,306 -/- 1,400 7 2,300 42 2,305 -/- 2,100 7 2,298 49 2,304 -/- 2,940 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters With a revenue of one full scalp of EUR 197.50, the losses from accumulated positions will start to dwarf the scalp revenue 1,000 2,310 2,308 2,306 2,304 2,302 2,300 2,298 2,296 2,294 2,292 (1,000) (2,000) (3,000) position loss one scalp (4,000) (5,000) (6,000) (7,000) As the position accumulates, please note that 14 points (60 bps) already generates a loss of almost USD 4,000, whereas the revenue of one scalp is about 1/20 thereof… 107 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters Left unchecked, our position continues to grow if the trend persists with position losses accelerating At one point our position is so large that revenue from one scalp is dwarfed by the impact of one tick movement in the future One could consider applying time-based position adjustment as well “if the trend does not reverse within 2 minutes, my original assumption of immediate mean-reversion is flawed, thus I need to lower the theoretical value of the EURO STOXX 50 by 1 point” 108 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Defining trading parameters The parameters as defined in the previous slides are crucial to the success of a pairs trading strategy: • Credit: a wider credit will make for more profitable scalps, though less frequent • Position Adjustment: a more aggressive position adjustment will reduce the severity of accumulation, but at the cost of reducing scalp revenue • Time-based Adjustment: quicker reduction of position will further reduce the severity of accumulation, but temporary losses will be locked-in more quickly => Set the right parameters (and recognize the regimes for each set) ! 109 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Backtesting and Risk Tolerance => Set the right parameters (and recognize the regimes for each set) ! Easier said than done, but l’histoire se repete (and we have buckets of historical data) 110 - What pair to select: * correlation and also co-integration * short-term volatility exceeds long-term volatility - What settings to use: * run all varieties though the historical data-set - Regimes: * higher volatility would call for larger credits * Increased trendiness would call for more aggressive adjustments © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K Mechanics of a spread trading strategy: Technical Dependencies •Feed: we quote in Stoxx based on the DAX; if our pricefeed for the DAX is delayed, we will quote incorrect prices •Position : we base our adjustments on our position; if we receive incorrect position data, we will quote incorrect prices •Latency: the products are correlated; if we sell in the Stoxx, we need to be fast in getting our hedge trade in the DAX •Trading state: we assume both markets are trading simultaneously; what happens if a circuit breaker triggers an auction (CAC vs DAX !) 111 © True Partner Education Ltd 2014 Spreading two correlated instruments C U H K And finally a word of warning: Mind the tail events! 112 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Hang Seng H-Shares Contract Size HKD 50 per point HKD 50 per point Typical Spread Width 1.00 point 2.00 points Last Traded 20,000 10,000 Contract Value HKD 1,000,000 HKD 500,000 Spread (bps) 0.5 bps 2.0 bps How would we trade this pair: a) Quote in Hang Seng with 1 future, hedge in H-Shares with 2 futures b) Quote in Hang Seng with 2 futures, hedge in H-Shares with 1 future c) Quote in H-Shares with 2 futures, hedge in Hang Seng with 1 future d) Quote in H-Shares with 1 future, hedge in Hang Seng with 2 futures 113 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Let’s we apply a ratio of 0.50 and require a credit of 2 points If we observe the following markets: Bid Price Ask Price Hang Seng 20,002 20,003 H-Shares (market) 10,002 10,004 a) We will join the offer at 10,004 points b) We will improve the offer to 10,003 points c) We will hit the bid at 10,002 points 114 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Let’s we apply a ratio of 0.50 and require a credit of 2 points If we observe the following markets: Bid Price Ask Price Hang Seng 20,002 20,003 H-Shares (market) 10,002 10,004 H-Shares (our) 9,999 10,004 So if you would look in terms of the ratio between Hang Seng and H-Shares, we are quoting the following Bid price: 0.4999 (9,999 vs 20,002) Ask price: 0.5001 (10,004 vs 20,003) 115 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Let’s we apply a ratio of 0.50 and require a credit of 2 points If we observe the following markets: Bid Price Ask Price Hang Seng 20,002 20,003 H-Shares (market) 10,002 10,004 H-Shares (our) 9,999 10,004 Assume we would do a full ‘scalp’ (i.e. buying and selling H-Shares at our prices and hedging in the Hang Seng, how much do we earn: a) b) c) d) 116 HKD 200 HKD 450 HKD 500 Cannot calculate © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Let’s we apply a ratio of 0.50 and require a credit of 2 points If we observe the following markets: Bid Price Ask Price Hang Seng 20,002 20,003 H-Shares (market) 10,004 10,006 H-Shares (our) 9,999 10,004 Most likely, we would start trading with one side: a ‘half scalp’ As we do not want to limitlessly increase our position, we need to make an adjustment after having traded: “for every 2 H-Shares futures I buy (sell), I will decrease (increase) my theoretical value by 2 points” 117 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Our pricing formula now becomes: - Apply a ratio of 0.50 - Apply a credit of 2 points per trade - Apply a position adjustment of 2 points per 2 futures (trade size) If we observe the following markets: Bid Price Ask Price Hang Seng 20,002 20,003 H-Shares (market) 10,004 10,006 Assuming our position is short 2 futures, our new market will become: a) 9,997 bid at 10,002 b) 9,999 bid at 10,004 c) 10,001 bid at 10,006 118 © True Partner Education Ltd 2014 Creating a Future Pairs strategy C U H K The Hang Seng vs H-Shares futures Our pricing formula now becomes: - Apply a ratio of 0.50 Apply a credit of 2 points per trade Apply a position adjustment of 2 points per 2 futures (trade size) While I reduce the risk of accumulating positions, I have not eliminated such risk: if the H-Shares keeps trending higher we will still increase our position Suppose we do the following sequence of trades in both products: - sell 2 HHI at 10,004 sell 2 HHI at 10,006 sell 2 HHI at 10,008 sell 2 HHI at 10,010 sell 2 HHI at 10,012 & & & & & buy 1 Hang Seng at 20,003 buy 1 Hang Seng at 20,003 buy 1 Hang Seng at 20,003 buy 1 Hang Seng at 20,003 buy 1 Hang Seng at 20,003 Based on the last traded prices (10,012 and 20,003), what is my result? 119 © True Partner Education Ltd 2014 Creating a Future Pairs strategy The Hang Seng vs H-Shares futures C U H K Our pricing formula now becomes: - Apply a ratio of 0.50 Apply a credit of 2 points per trade Apply a position adjustment of 2 points per 2 futures (trade size) Apply a time-based adjustment of 1 point per minute Right now, I have the following position: Short 10 HHI futures (last sale at 10,012) Long 5 Hang Seng futures (last purchase at 20,003) Bid Price Ask Price H-shares (our) 20,002 20,003 H-Shares (market) 10,012 10,014 Applying the time-based adjustment, how long will it take before we have unwound the entire position? a) b) c) d) 120 10 minutes 11 minutes 12 minutes 13 minutes © True Partner Education Ltd 2014