UANT : OB Orestes : Answersbesa Rce ea TOES ELE CL Letsn ta PTET acest) eet leT eee) Priiccagonms Monel om iO Caate | interviews in the City and Wall Street. bere me aarormcen oma ew atl meccrel Coed Rots Cera SCOUR AMICUS Laas Clk Pears retirees te Ores och Co) Certo aaa (On Topics covered include option pricing, Pucci amen nie onnteremeE Caste algorithms and C++, as well asa discussion nd the of the interview proce: Pre Ecce atovET cast and VEY alos Tacs a un Cre nec cco ay textbooks “the Concepts and Practice of Mathematical Finance” and “C++ Design Patrerns and Derivatives Pricing.” BS Con oe cote CE Roe Ohne cm elniaag POSUE UNA TE Ue Cy y onamee ccs (ale ocnae tates ironed ts leRe meter aa S NavB-ex807709 i 7031Quant Job Interview Questions and Answers Mark Joshi Nick Denson Andrew Downes@©Mark Joshi, Nick Denson, Andrew Downes 2008 Version 1.01Contents Preface ‘What this book is and is not How to use this book Website Structure The publication of this book Chapter 1. The interview process 1.1. Introduction 1.2. Getting an interview 1.3. The standard interview 1.4. The phone interview 1.5. The take-home exam 1.6. The exam 1.7. Follow-up 1.8. Dos and don'ts 1.9. When to apply? 1.10. The different roles LIL. Sorts of employers 1.12. Where people go wrong Chapter 2. Option pricing 2.1. Introduction 2.2. Questions 2.2.1. Black-Scholes 2.2.2. Option price properties 2.2.3, Hedging and replication 2.2.4. The Greeks viii viti ix 15 15 16 16 7 19 20wv CONTENTS 2.25. General 2.2.6. Trees and Monte Carlo 2.2.7. Incomplete markets 23. Solutions 2.3.1. Black-Scholes 2.3.2. Option price properties 2.3.3. Hedging and replication 2.3.4, The Greeks 23.5. General 2.3.6. Trees and Monte Carlo 2.3.7. Incomplete markets Chapter 3. Probability 3.1. Introduction 3.2. Questions 3.2.1. General 3.2.2. Stochastic processes 3.3. Solutions 3.3.1. General 3.3.2. Stochastic processes Chapter 4. Interest rates 4.1, Introduction 4.2. Questions 43. Solutions Chapter 5. Numerical techniques and algorithms 5.1. Introduction 5.2. Questions 53. Solutions Chapter 6. Mathematics 6.1. Introduction 6.2. Questions 6.2.1. General 6.2.2. Integration and differentiation 6.3. Solutions 2 21 22 22 33 51 58 62 65 R 19 RESS 85 124 147 147 148 148 157 157 157 159 189 189 189 189 191 192CONTENTS 6.3.1. General 6.3.2. Integration and differentiation Chapter 7. Coding in C++ 7.1. Introduction 7.2. Questions 7.3. Solutions Chapter 8. Logic/Brainteasers 8.1. Introduction 8.2. Questions 8.3. Solutions Chapter 9. The soft interview 9.1. Introduction 9.2. Soft questions and answers 9.3. Finance data questions Chapter 10. ‘Top ten questions 10.1, Introduction 10.2. Questions Bibliography Index 192 210 223 223 226 233 215 215 216 279 297 297 207 303 305 305 305 307 309Preface What this book is and is not ‘The purpose of this book is to get you through your first interviews for quant jobs. We have gathered a large number of questions that have actually been asked and provided solutions for them all. Our target reader will have already studied and learnt a book on introductory financial mathematics such “The Concepts and Practice of Mathematical Finance.” He will also have learnt how to code in C++ and coded up a few derivatives pricing models, and read a book such as “C++ Design Patterns and Derivatives Pricing.” ‘This book is not intended to teach the basic concepts from scratch, instead it shows how these are tested in an interview situation. However, actually tackling and knowing the answers to all the problems will undoubtedly teach the reader a great deal and improve their performance at interviews. Many readers may find many of the questions silly and/or annoying, so did the authors! Unfortunately, you have to answer what you are asked and thinking the question is silly does not help. Arguing with the interviewer about why they asked you it will only make things worse. For that reason, we have included many questions which we would never ask and think that no one should ask. ‘That said, if the questions are too silly then you may want to consider whether you want to work for the interviewer. How to use this book We strongly advise you to attempt the questions seriously before looking at the answers. You will learn a lot more that way. You may also come up with different solutions. We have included a number of follow-up questionsvit PREFACE without solutions in the answers, which may or may not have been asked in a live interview. Tackling these will help you refine your skills. Some questions come up time and time again, so if you have actually learnt all the answers then there is no doubt you will eventually get some duplicates, but that is not really the point; it is more important to be able to tackle all the types of questions that arise and to identify your weaknesses so you can address them. Eventually, of course, the interviewers will buy copies of this book to make sure that the questions they use are not in it. However, in the meantime, you can make the most of your comparative advantage. Many candidates at their initial interviews seem to have a poor idea of what is required. If you find the questions in this book unreasonably hard, you are not yet ready. If you think that you will never be able to do them, then now is a good time to think of an alternative career. If you are on top of these sorts of questions, you should have no problem getting an entry level job. So use this book to judge when you are ready. If possible, get a friend, who is experienced, to give you a practice interview when you think you have reached that point. Website Inevitably, readers will have plenty of questions regarding this book’s contents. For that reason, there is a forum on www.markjoshi .com to discuss its content. In particular, if you think a solution is wrong, or want a solution to a follow-up. question, then this is the place to ask. There are also a lot of additional resources such as recommended book lists, discussion of problems, job adverts and career advice on that site. Structure This book is structured as follows: we start with a discussion of the interview process in Chapter 1 including how to get one as well as how to conduct yourself during one. We then move on to actual interview questions; each chapter contains some general discussion, a set of questions from real-life interviews, and then solutions with follow-up questions. We divide these into topics by chapter: option pricing, probability, interest rates, numerical techniques, mathematics, coding in‘THE PUBLICATION OF THIS BOOK ie C++, and brainteasers. We then discuss how to handle a “soft interview”, that is a non-technical interview, and list possible questions. We finish with a list of ten, of the most popular questions from quant interviews. The publication of this book ‘The reader may be curious to know how this book has been published. We decided not to go with conventional publishers for a number of reasons. The first is simply that the authors lose all control; once you have signed on the dotted line all the understanding and reassurances from your editor become worthless. ‘A second is timing, a conventional publisher can easily take two years to get a book from draft to release. We have therefore gone with print-on-demand direct sales. With this approach to publishing, the book is printed and bound after the order is made. There is therefore no overhead but the final product is of equally high quality. We hope that you will support this publishing revolution with your purchases and the books you write.CHAPTER 1 The interview process 1.1. Introduction In this chapter, we look at how to get a job interview and then what happens once you have one. We also discuss the types of interview, the general process and what happens afterward. It is important to realise that how you behave in and approach the interview can have a marked effect on your chances. We round off with a discussion of the different roles, areas and types of employers. There are, of course, many sorts of interviews and many ways for the -wer to conduct it, but ultimately the interviewer wants to find out two © Do you have the technical ability to do the job? © Will you behave reasonably? That is, will you do what you are asked and get on with others. The first will be assessed via a barrage of technical questions, and the second by observing your behaviour and on how you respond to general questions. Most interviews focus largely on technical ability, since most candidates fail at this point. However, on the rare occasion there are multiple candidates who are good enough technically, other factors do become important. 1.2. Getting an interview For many candidates, the biggest problem is getting the interview in the first place. There are many ways to get that far. The first thing to remember is that in hiring “like attracts like.”1. THE INTERVIEW PR Do you know anyone with a similar background who's now working in the City? If so, ask if you can come down for a chat. This may translate rather quickly into a job offer. If a PhD student the year above you is applying for jobs in finance, make sure to make friends with him before he leaves. Keep in touch and when the time comes make use of the contact. Once you have exhausted all the people you know with similar backgrounds, try the ones with different backgrounds and then friends of friends. Some places, for example BarCap, now have quantitative associates. pro- grammes and you apply via the bank’s web-site, otherwise interviews tend to come via recruitment consultants also known as “headhunters.” Some recruitment firms now even map out the PhD classes in mathematics and physics departments at top universities: if you are looking for someone to sponsor your social events, they are a good place to start! Headhunters generally call you down for a meeting to make sure that you are presentable. They then send you to a couple of interviews to see how you do. If you do well they will get very enthusiastic and then send you to lots of places. If you do badly they will quickly lose interest and it is time to find a new headhunter. A much-discussed and difficult to answer question is how many firms to use at once. My inclination is to start with one or two and see how it goes. As Jong as they are getting you plenty of interviews, there is no point in registering somewhere else. But if they aren't, then it is time to try another firm. ‘An important fact to realise is that headhunters are paid by commission (e.g. 20% of your first-year package) on a placement basis. The motivations and incentives are therefore a lot like real estate agents: « they really want to place you (good); « they want you to get a high salary: commission is a percentage (good); « they would much rather place you quickly in any job than slowly in a better one (not so good); © they will be unhappy if you accept a job through someone else (no commission 10 fee == very unhappy); ‘ they may or may not be interested in a long term relationship; you will be worth five times as much in five years to them,13. THE STANDARD INTERVIEW 3 A headhunter’s business is relationships and information. They are therefore keen not to damage their reputation and also to get as much information as possible about who you know and who is doing what. Always bear this is mind when talking to them. Once you actually get a job, you will regularly get calls from them fishing for information and seeing if you are interested in roles they won't be keen to tell you much about until you are firmly their client. It is OK to talk to them but do NOT under ANY circumstances discuss any of your colleagues; this could get you fired. (For similar reasons, never ever talk to journalists but refer them to corporate communications or your boss.) ‘Two headhunter firms that have their own guides to becoming a quant and have good reputations are: # Paul and Dominic. The “Paul” is Paul Wilmott the financial mathemati- cian. They have the advantage of actually understanding the job they are trying to place you in. © Michael Page. 1.3, The standard interview ‘The most common interview consists of arriving at the bank’s offices, you ask for some person whose name you have been given, you sit in the lobby and wait, Eventually someone comes and gets you, you are then shown into a meeting room. There may be one or several interviewers present. You have already had several chances to mess up. Make sure that: '* you arrive dead on time: being early really irritates, and being late displays disorganisation; * you wear a suit and are looking well groomed; you know and remember the name of the person you are meeting; ‘* you have a copy of your CV with you, and do not expect them to have seen nor read it; ‘you have had plenty to eat and are not suffering a sugar low since that, will destroy thinking power; ‘* you switch off your mobile phone.4 1. THE INTERVIEW PROCESS Arriving dead on time is an art: the only way to do it is to arrive with half an hour to spare, and then go and find a cafe where you can have something to eat or drink. Having a lemonade is a good idea to keep sugar levels up. Don't drink too much, however, or you will be rushing to the loo in the middle of the interview. Also try to assess how long the queue at reception is. If it is looking Jong you may need to join it ten minutes before your scheduled arrival time. ‘The interviewer may or may not be the person who fetched you from reception. ‘When the interviewer arrives he (or she) will typically give you the chance to ask a few questions. Whilst it is good to ask a little, e.g. what the team does, where it sits, how many people are in it, what sort of role you are being considered for, it is best not to drag this part out since if you do badly in the technical interview it is all rather irrelevant. In addition, the interviewer may find too many questions at this stage annoying. There are often two or more interviewers rather than one, one will typically take the lead, however. You will then be asked very technical questions and typically be given either a whiteboard or a sheet of paper to work through them. If you get stuck, the interviewer will generally help you out, the more help you need the worse you have done. Some interviewers always ask the same level of questions, others will make the questions harder if you get them right, and easier if you get them wrong, Sometimes the interviewer will vary the questions to try and find what you are good at, if anything. Important points to remember are: ‘* Do not argue with them about why they asked something. If they asked, they want the answer. Disputing will make you look difficult and weak technically. ‘© The thinking process counts as well as the solution, so talk about how you are tackling it ‘© They do not expect you to be able to do everything without help. If you are unsure of what they want, ask for clarification. For example, for a numerical or coding problem, do they just want a short solution, or do they want an optimal one? ‘Their expectations of you will be affected by your c.v. For example, if you say that have read the interviewer's book then you had better know it well. (Authors4. THE PHONE INTERVIEW 5 are very vain about their books so do not say anything critical about it either.) If you say you are an expert in a coding language, e.g. C++, then be sure you can back the statement up. [At the end, if you have done poorly, accept this graciously and ask for advice on what to read and how to improve. Do not get into an argument with the interviewer about what happened; just notch it up to experience and resolve to do better next time, Banks in the UK tend not to bother with paying expenses and particularly not for first interviews. In the US, they will sometimes pay depending upon the length of the trip. If you know that you will be there all day, it is worth taking some refreshments in your bag. Sugar lows and dehydration will badly affect your ability to think. Don't assume that it will have occurred to them that you need to eat and drink, just because you are there from nine to five. Goldman Sachs is most notorious for grilling you across long days by several people. They really do want you to meet everyone and if you cannot take the pace of the interviews, you won't be able to cope with working there. 1.4. The phone interview Phone interviews are generally used when it is not convenient to see the candidate in person, They are often a pre-screen to check whether a proper interview is worth the effort when travel is involved. When a team is international then they are used by the overseas part of the team. You will never get a job without a face-to-face interview so this can only ever be one stage of the process. ‘The process is not hugely different from an in-person interview. Generally, there is only one interviewer and they ask you a series of questions and chat with you a little. The nature of the questions tends to be a little different since they need to work over the phone. For that reason, the fraction of brainteasers are higher than normal.6 1. THE INTERVIEW PROCESS ‘The important points to remember are: use the best phone line you can, (which is generally not a mobile;) speak loudly and clearly; # do not give the impression that you are looking things up in a book or on the web have pen and paper handy. 1.5. The take-home exam ‘This again is used to decide whether it is worthwhile to bother with a proper interview. The questions will largely reflect the interests of the hiring manager so they may focus on one tiny aspect of applied mathematics or probability theory or even physics. Generally, they will e-mail it to you and expect a response within some set time period, e.g., 24 hours. Do not try to negotiate the time available to do it, since this gives the impression that you are weak mathematically and generally difficult. It is acceptable to ask that the start time be moved, e.g., to a Saturday so that you can devote yourself to it If you decide to copy out solutions from a book, try not to be too blatant and make sure you copy out the right material. One candidate copied out the code for a vanilla call option from the interviewer's book when asked to code up an Asian option pricer. His application was not taken further, ‘The presentation of the answers matters as well as correctness. So make sure your handwriting is clear and your steps are clearly explained, 1.6. The exam The written exam which is not take-home is becoming more popular. Some banks are even setting a general exam for a large number of candidates at once and then taking the highest performers. This has the virtue of clarity and fairness. It also favours people who are good at exams rather than interviews, for better or worse.17. FOLLOW-UP 7 The main problem is that the questions tend to reflect the background of the setter rather than relevance to the job. However, this is no different from interviews. If you know you will be doing a written exam, then find out what the rules are. For example, is it open book? How long will it last? Are you allowed to use a calculator? Make sure to bring your own calculator ~ sitting in a room on your own with a defective calculator (or none at all) can be very stressful. 1.7, Follow-up Most places will not give you feedback on the spot, but some occasionally will. If you got the interview via a recruitment agent, (ie., headhunter) it will generally come that way. Otherwise, expect an e-mail a few days afterward. If you do not hear anything for a week or two, then it is perfectly reasonable to politely ask what is happening. If you have done well, they may move very quickly since very few really good candidates come along. Bear in mind, that if they do not want you for the job, it does not mean that they think poorly of you. The first author of this book is in touch with quite a few people he met when turning them down for a job ~ often it just means that the preparation was not quite right or there was a better candidate. These candidates showed some potential and are now leading successful city careers after taking the feedback that they were given seriously. Related to this, remember that the quant circle is not very big so you will come across the same people repeatedly — don’t destroy any relationships unnecessarily. Indeed, it is not unusual to find that after a takeover, you are working for the person who rejected you a year before. It is very rare to get the job after the first interview. Instead, if you do well they will invite you back to meet more people. If they are organised and keen ‘you may have several interviews in one day, or they may get you back again and again, until you have met everyone. After two or three rounds of this they will make a decision. Generally the more times you are invited back, the keener they are. However, it can just reflect an inability to pool information. Going to ten interviews and then being told you are not getting the job because the second interview went badly is not unknown.|. THE INTERVIEW PROCESS 1.8, Dos and don’ts Here’s a checklist of things to do and not do: don't be late; don’t be early; don’t argue with the interviewer about why they've asked you something; do appear enthusiastic; do wear a suit; do be eager to please (they want someone who will do what they want, you must give the appearance of being obliging rather than difficult); © don’t be too relaxed (they may well conclude that you aren’t hungry enough for success to work hard); © don’t tell them they shouldn’t use C++ because my niche language is better, ¢ do demonstrate an interest in financial news; do be able to talk about everything on your CV ( also known as resumé) have a prepared 2 minute response on every phrase in i ‘* do bring copies of your CV; © don’t expect the interviewer to be familiar with your CV; © don’t say you've read a book unless you can discu: particularly if they’ve written it; *# do be polite; ¢ do ask for feedback and don't argue about it (even if it is wrong try to understand what made the interviewer think that); ‘© don't say you want to work in banking for the money (of course you do, but it’s bad form to say so); ‘* do say you want to work closely with other people rather than solo; ‘© don’t say that you think that bankers are reasonable people — they aren't; © do take a break from interviewing and do more prep if more than a couple of interviews go badly; * don’t use a mobile for a phone interview; * do be able to explain your thesis ~ work out explanations for different sorts of people in advance; © don’t expect banks in the UK to pay for interview expenses; # don’t spend more on expenses than has been agreed; its contents;9. WHEN TO APPLY? ° ‘ do ask about the group you'll be working in: ~ how much turnover is there? — where people go when they leave? ~ how many people are in the team? = when can you meet the rest of the group? (only ask this if an offer appears imminent; if you can’t meet the others, this is a big red flag: what's wrong with them?) — how old the group is? — what’s the team’s raison d'etre? ~ is it expanding or contracting? = what would a typical working day be? ‘¢ don’t get on to the topic of money early in the process; # don’t be cynical about what bankers do; ¢ don’t accept an offer made under pressure. 1.9. When to apply? Most entry-level quants are hired because a specific team has a need for someone. This means that they want someone to start now and they want someone who will be productive quickly. This means that you should not start applying unless you can start within the next two months, You should also wait until you are well-prepared. This is doubly the case in smaller places. In London or New York, you can learn the hard way that you are not ready, but in Melbourne or somewhere similar, there may be only two or three possible employers so you had better be sure you perform at your best from the start. How can you tell if you are ready? Here are some indicators: ‘© Could you get “A” in an exam on the contents of “the Concepts and Practice of Mathematical Finance” [6] ? ‘© Have you coded up some models in C++? (e.g. the computer projects at the end of that book.) ‘® Are you on top of the contents of “C++ Design Patterns and Derivatives Pricing” [7] ?10 |. THE INTERVIEW PROCESS © Can you do the interview questions in this book without too much difficulty? # Can you tackle the supplementary questions in this book? © Have you successfully completed a practice interview? The rules that apply to quantitative associates programs are different, since they will generally only be open at one point in the year. In these cases, find out what they want and what flexibility they have. Also, find out if failing one year will count against you the next year and take that into account too. 1.10. The different rotes It is important to realise that there are many different types of quants who do different sorts of things. There are pros and cons of each and it is worth considering what sort of role you want, and communicating that to potential employers. A brief list is: (1) front office/desk quant; (2) model validating quant; (3) research quant; (4) quant developer; (5) statistical arbitrage quant; © capital quant; (7) portfolio theorist. ‘A desk quant implements pricing models directly used by traders. This can mean either very short term projects or longer term ones depending on the way the outfit is setup. The main advantage is that you are close to the real action both in terms of things happening and in terms of money. This is also a possible route into trading. The downside is that it can be stressful and depending on the outfit may not involve much research. A model validation quant independently implements pricing models in order to check that front office models are correct. It tends to be more relaxed, and less stressful. The downsides are that model validation teams can be uninspired and far from the money. Also in some places, the quants spend their time running other peoples’ models rather than coding their own which can be quite frustrating.1.10. THE DIFFERENT ROLES n A research quant tries to invent new pricing approaches and sometimes carties out blue-sky research, These are the most interesting quant jobs for those who love mathematics, and you learn a lot more. The main downside is that it sometimes hard to justify your existence. Quantitative developers are programmers who generally implement other people’s models. It is less exciting but generally well-paid and easier to find a job. This sort of job can vary a lot. It could be coding scripts quickly all the time, or working on a large system debugging someone else’s code. The statistical arbitrage quant works on finding patterns in data to suggest automated trades. The techniques are quite different from those in derivatives pricing. This sort of job is most commonly found in hedge funds. The return on this type of position is highly volatile! A capital quant works on modelling the bank’s credit exposures and capital requirements. This is less sexy than derivatives pricing but is becoming more and more important with the advent of the Basel II banking accord. You can expect decent (but not great) pay, less stress and more sensible hours. There is currently a drive to mathematically model the chance of operational losses through fraud etc, with mixed degrees of success. The biggest downside of going into this area is that it will be hard to switch to derivatives pricing later on. Portfolio theorists use financial mathematics in the sense of Markowitz’s portfolio theory rather than derivatives pricing and Black-Scholes. It is less technically demanding but there is certainly plenty of money in the area. There is a certain commonality between this area and capital modeling. Again it is hard to switch from this to derivatives pricing. People do banking for the money, and you tend to get paid more the closer you are to where the money is being made. This translates into a sort of snobbery where those close to the money look down on those who aren't, As a general rule, moving away from the money is easy, moving toward it is hard.2 |. THE INTERVIEW PROCESS 1.11. Sorts of employers There is quite a lot of variety in terms of sorts of employers. We give a rough catalogue: © commercial banks, e.g., RBS, HSBC; © investment banks, ¢.g., Goldman Sachs, Lehman Brothers; @ hedge funds, e.g., the Citadel Group; -countancy firms; software companies. Each of these has its pros and cons. Large commercial banks tend to have large trading operations, but are influenced by the culture of the rest of the bank. The effect of this is that they tend to be less tough but also less exciting in terms of products and projects compared to investment banks or hedge funds. The advantages are shorter hours and better job security. The main disadvantage tends to be less money! Investment banks, particularly American ones, tend to expect longer hours and have a generally tougher culture. They are much readier to hire and fire. If you want an astronomical bonus, however, they are the place to go. Hedge funds tend to demand a lot of work. They are very volatile and have been a big growth industry in recent years. They have, however, been badly hit by the credit crisis in 2008; they may, or may not, emerge well. The packages tend to reflect very large risk premia. In general, American banks and firms pay better but demand longer hours than European ones. The big accountancy firms have quant teams for consulting. The main disadvantage is that you are far from the action, and high quality individuals tend to work in banks so it may be hard to find someone to learn from. Some firms are very good on external employee training, however, and will send employees ‘on Masters courses or regular training courses. There is an increasing move towards outsourcing quant modeling by pur- chasing off the shelf software models. One option is therefore to work for the software company instead. The issues are similar to those with working for1.12, WHERE PEOPLE GO WRONG 3 accountancy firms. The growth in availability of open-source financial software such as QuantLib may hit these companies in the medium term. 1.12. Where people go wrong A certain number of people try and fail to get quant jobs; it therefore has a reputation as a tough area to get into. The biggest reasons for failure are: mistaken ideas about the knowledge required; inability to code; personality defects; non-possession of appropriate degrees; lack of ability at mathematics; misperception of own ability. eee eoe If none of these apply, and you have done your preparation then it is actually quite easy to get a job, and you will be snapped up within weeks if not days. What if some do apply? This book should make it clear what is required and how to acquire the necessary knowledge. If you don’t know how to code, then you simply have to learn by picking up the books and coding some models. If you can't do this, try a different career. If you have personality defects then well done for recognising the fact. Quant jobs are not an area where personality counts for a lot at entry level. Try reading, a few books in the “self-help” section of the book-shop and work on your people skills. You only have to appear normal for a couple of hours to get the job. The simple truth is that if you apply for quant jobs without something that says you are really good at maths in your c.v. then you won't get interviews. You therefore have to get a degree that demonstrates the ability and knowledge you claim, or do something that shows the requisite skills in other ways. If you simply aren’t that great at mathematics then this is not the career for you. Even if you manage to get that first job, you will be working day in, day out with people who love mathematics and can’t imagine doing anything else. You will not thrive in that environment, better to find something you have natural talent in,“ |. THE INTERVIEW PROCESS Inability to assess one’s own ability quickly shows up when you think you are on top of everything and you start bombing the interviews. If you keep failing the interviews, it is a strong lesson that you need to reassess yourself. Getting a friend who is already in the quant area to do a practice interview is a good way to assess your ability. Working through the problems in this book and seeing how many you can do without help is another way. Ultimately, there aren’t really that many people who are very strong at mathematics. If you are one of those people, the quant career is very attractive and easy to get into; if you aren’t then it is hard,CHAPTER 2 Option pricing 2.1. Introduction The majority of work for quants in a bank is focused on the pricing of options. It is not surprising then that a large section of this book is dedicated to option pricing questions. Before even looking at financial models however, one needs to understand some of the more fundamental properties of option prices, such as no arbitrage bounds. For example, how does the price of a call option vary with time? What happens as volatility tends to infinity? The classical model of Black and Scholes is almost certain to come up in any interview, so make sure you understand this model. You should be able to derive the pricing formula for at least a European call option and be able to extend it to different payoffs. It is also worth understanding the Greeks: what they mean and what they are in the Black-Scholes model. Another key aspect of financial modeling is hedging and replication. Having a good understanding of what replication is and how you can replicate an unusual payoff with vanilla options is a valuable skill. Some banks manage to make large sums of money by replicating an exotic option with vanilla options, and you will be expected to have a good understanding of replication: both static and dynamic. We briefly mention a few introductory books on option pricing. We also refer the reader to a much longer list, which is occasionally updated, on www markjoshi .com. This book of interview questions can be viewed as a companion book to the first author’s book on derivatives pricing: “The Concepts and Practice of 156 2. OPTION PRICING Mathematical Finance.” For those who wish to have some alternatives, here are some standard choices: J. Hull, “Options, Futures and Other Derivatives,” — sometimes called the “bible book.” Gives a good run-down of how the markets work but is aimed at MBAs rather than mathematicians so the mathematics is quite weak. © T. Bjork, “Arbitrage Theory in Continuous Time.” This book is on the theoretical side with the author having a background in probability theory, but he also has a good understanding of the underlying finance and he is good at translating intuition into theory and back. © S. Shreve, “Stochastic Calculus for Finance Vols I and I.” A careful and popular exposition of the theory. * P, Wilmott, various books. Good expositions of the PDE approach to finance, but not so good on the martingale approach. ¢ M. Baxter and A. Rennie, “Financial Calculus.” A good introductory book on the martingale approach which requires a reasonable level of mathematical sophistication but also has good intuition. 2.2. Questions 2.2.1, Black-Scholes. Question 2.1. Derive the Black-Scholes equation for a stock, S. What boundary conditions are satisfied at § = 0 and S = co? Question 2.2. Derive the Black-Scholes equation so that an undergrad can understand it. Question 2.3. Explain the Black-Scholes equation. Question 2.4. Suppose two assets in a Black-Scholes world have the same volatility but different drifts. How will the price of call options on them compare? Now suppose one of the assets undergoes downward jumps at random times. How will this affect option prices? Question 2.5. Suppose an asset has a deterministic time dependent volatility. How would I price an option on it using the Black-Scholes theory? How would T hedge it?22. QUESTIONS. ” Question 2.6. In the Black-Scholes world, price a European option with a payoff of max(S} — K,0) at time T. Question 2.7. Develop a formula for the price of a derivative paying max(Sr(Sr — K), 0) in the Black-Scholes model. 2.2.2. Option price properties. Question 2.8. Sketch the value of a vanilla call option as a function of spot. How will it evolve with time? Question 2.9. Is it ever optimal to early exercise an American call option’ What about a put option? Question 2.10. In FX markets an option can be expressed as either a call or a put, explain, Relate your answer to Question 2.9. Question 2.11. Approximately how much would a one-month call option at-the-money with a million dollar notional and spot 1 be worth? QuesTion 2.12. Suppose a call option only pays off if spot never passes below a barrier B. Sketch the value as a function of spot. Now suppose the option only pays off if spot passes below B instead. Sketch the value of the option again. Relate the two graphs. Question 2.13. What is meant by put-call parity? Question 2.14. What happens to the price of a vanilla call option as volatility tends to infinity? Question 2.15. Suppose there are no interest rates. The spot price of a non-dividend paying stock is 20. Option A pays 1 dollar if the stock price is above 30 at any time in the next year. Option B pays 1 if the stock price is above 30 at the end of the year. How are the values of A and B related? Question 2.16. How does the value of a call option vary with time? Prove, your result.8 2. OPTION PRICING Question 2.17. A put and call on a stock struck at the forward price have the same value by put-call parity. Yet the value of a put is bounded and the value of a call is unbounded. Explain how they can have the same value. Question 2.18. Suppose we price a digital call in both normal and log- normal models in such a way that the price of call option with the same strike is invariant. How will the prices differ? Question 2.19. What is riskier: a call option or the underlying? (Consider a ‘one-day time horizon and compute which has bigger Delta as a fraction of value.) Question 2.20. If the stock price at time T’ is distributed as N(So, 0?) what is the expected value of an at-the-money European call expiring at time T? Question 2.21. Assume that the price of a stock at time T’ is N(So, 0) where So is the price now and that we know the price of an at-the-money European call expiring at T. How could we estimate 0? Question 2.22. A stock $ is worth $100 now at t = 0. At t = 1, S goes either to $110 with probability = 2/3 or to $80 with prob 1/3. If interest rates, are zero, value an at-the-money European call on expiring at ¢ = 1. Question 2.23. Sketch the value of a vanilla call and a digital call as a function of spot. Relate the two. Question 2.24. Price a1 year forward, risk free rate = 5%, spot = $I and a dividend of $0.10 after 6 months. Question 2.25. What is the fair price for FX Euro/dollar in one year? Risk five rates and spot exchange rate given. Question 2.26. An option pays 1, if > S2, {° otherwise, at time T- If the volatility of 5; increases, what happens to the value of the option? Question 2.27. In the pricing of options, why doesn’t it matter if the stock price exhibits mean reversion? Question 2.28. What are the limits/boundaries for the price of a call option on a non-dividend paying stock?22. QUESTIONS 19 Question 2.29. What is the value of a call option for a 98th percentile fall in the stock price? Question 2.30. What is the price of a call option where the underlying is the forward price of a stock? Question 2.31. Prove that the price of a call option is a convex function of the strike price. 2.2.3. Hedging and replication. Question 2.32. What uses could an option be put to? Question 2.33. Suppose spot today is 90. A call option is struck at 100 and expires in one year. There are no interest rates. Spot moves log-normally in a perfect Black-Scholes world. I claim that I can hedge the option for free. Whenever spot crosses 100 in an upwards direction I borrow 100 and buy the stock. Whenever spot crosses 100 in a downwards direction I sell the stock and repay my loan. At expiry either the option is out-of-the-money in which case I have no position or it is in-the-money and I use the 100 dollar strike to payoff my loan. Thus the option has been hedged for free. Where is the error in this argument? Question 2.34. ‘Team A plays team B, in a series of 7 games, whoever wins 4 games first wins. You want to bet 100 that your team wins the series, in which case you receive 200, or 0 if they lose. However the broker only allows bets on individual games. You can bet X’ on any individual game the day before it ‘occurs to receive 2X’ if it wins and 0 if it loses. How do you achieve the desired pay-out? In particular, what do you bet on the first match? Question 2.35. Suppose two teams play five matches. I go to the bookmakers and ask to place a bet on the entire series. The bookie refuses saying I can only bet on individual matches. For each match I either win X dollars or lose X dollars. How would I construct a series of bets in such a way as to have the same payoff as a bet on the series? Question 2.36. You want to bet $64 at even odds on the CWS winning the World Series. Your bookmaker will only let you bet at even odds on each game. What do you do?Ey 2. OPTION PRICING Question 2.37. Suppose an option pays 1 if the spot is between 100 and 110 at expiry and zero otherwise. Synthesize the option from vanilla call options. Question 2.38. Suppose an option pays zero if spot is Jess than 100, or pays spot minus 100 for spot between 100 and 120 and 20 otherwise. Synthesize the option from vanilla options. Question 2.39. What is pricing by replication? Question 2.40. Replicate a digital option with vanilla options. Question 2.41. The statistics department from our bank tell you that the stock price has followed a mean reversion process for the last 10 years, with annual volatility 10% and daily volatility 20%. You want to sell a European option and hedge it, which volatility do you use? Question 2.42. A derivative pays —— min(max(Sp, Ki), Ka)’ with Ky < Ka. Derive a model independent hedge in terms of a portfolio of vanilla options. 2.2.4, The Greeks. Question 2.43. What methods can be used for computing Greeks given a method for computing the price? What are their advantages and disadvantages? Question 2.44. How does the Gamma of a call option vary with time? Question 2.45. Suppose an option pays one if spot stays in a range Kl to K2 and zero otherwise. What can we say about the Vega? Question 2.46. All being equal, which option has higher Vega? An at-the- money European call option with spot 100 or an at-the-money European call option with spot 200? (A structurer asked this question and didn’t want formulas.) Question 2.47. How do you construct a Vega neutral portfolio with vanilla call and put options?2.2. QUESTIONS a 2.2.5. General. Question 2.48. How accurate do you think a pricing function should be? Question 2.49. Assume you have a good trading model that you think will make money. What information would you present to your manager to support your claim. 2.2.6. Trees and Monte Carlo. Question 2.50. A stock is worth 100 today. There are zero interest rates. The stock can be worth 90 or 110 tomorrow. It moves to 110 with probability p. Price a call option struck at 100. Question 2.51. At the end of the day, a stock will be 100 with probability p= 0.6 and 50 with probability 1 — p = 0.4. What is it trading for right now? Value an at-the-money European call option expiring at the end of the day. What if the actual stock price is 75 right now? Question 2.52. A stock is worth 100 today. There are zero interest rates. The stock can be worth 90, 100, or 110 tomorrow. It moves to 110 with probability p and 100 with probability q. What can we say about the price of a call option struck at 100. Question 2.53. Follow-up: given that we have seen that trinomial trees do not Iead to unique prices, why do banks use them to compute prices? Question 2.54, Consider the following binomial tree. There are two identical underlying assets A and B with the same prices and volatility. If all were the same except that research suggests company A will do better than company B, how would the option prices compare? Question 2.55. Monte Carlo versus binomial tree — when shall you use one or the other? Question 2.56. Current stock price 100, may go up to 150 or go down to 75. What is the price of a call option based on it? What is the Delta? Question 2.57. Explain the Longstaff-Schwartz algorithm for pricing an early exercisable option with Monte Carlo.2 2. OPTION PRICING 2.2.7. Incomplete markets. Question 2.58. What is implied volatility and a volatility skew/smile? Question 2.59. What differing models can be used to price exotic foreign exchange options consistently with market smiles? What are the pros and cons of each? Quesrion 2.60. Explain why a stochastic volatility model gives a smile. 2.3. Solutions 2.3.1. Black-Scholes. Solution to Question 2.1. In the Black-Scholes world the evolution of the stock price S; is given by dS, = wSidt + 05,dWi, for 1, 0 > 0. We also assume the stock does not pay any dividends, there are no transaction costs and the continuously compounding interest rate is r > 0 (constant). The latter assumption implies the evolution of the risk-free asset By is given by dB, = Bat. We are interested in pricing an option which is a function of the stock price at time T > 0, Sp. One possible example is a call option with strike K > 0, that is a derivative which at time T’ pays max(Sp — K,0). ‘While the form of the payoff is not particularly important, that it is a function of the stock price at time T, and only time 7, is important. Under this condition wwe can show that the call option price is a function of current time ¢ and current stock price S; only (see e.g. [16] p.267, Theorem 6.3.1 or [6]). Thus we denote by C(t, St) the call option price. To price a derivative in the Black-Scholes world, we must do so under a measure which does not allow arbitrage (clearly the existence of arbitrage in any model is cause for concern). Such a measure is called a risk-neutral measure.23. SOLUTIONS 2 One can prove that under this measure, the drift term of the stock price changes so that dS, = rS,dt + oS,dW,. We are now ready to proceed with our derivation. In the risk-neutral world, C(t, S:)/ Bz is a martingale and hence if we calculate its differential we know it must have zero drift. Applying It’s lemma to C(t, S:) gives ac, ac 18, ACU, 8) = Feat + Fg 4S + 3 Bop where the arguments of C’ and its partial derivatives are understood to be (t,5,). Using the risk-neutral dynamics of S; (and recalling that (dW)? = dt, dW,dt = (dt)? = 0) gives 10°C 4 62 aC(t, S,) = (F saa? ?) at 0S a 7 OS, Finally using the It6 product rule we can compute ‘C(t, Si) 1 (8C , OC, ,1PC 4.9 Sy OC a( a Bi (ae tag + 33g" —C) dtr oR eam. Since we know this is a martingale (the drift term is zero), we see that 6, AC 5 IAC ooo at t as," * 3ag77 (dS.)?, Wi 1) rO=0. This is the Black-Scholes equation, When considering the boundary conditions, we do need the form of the payoff function of the derivative. Here we take our example of the call option with strike A’. We can approach the question regarding the boundary conditions in two ways. The first is simple, logical, but not entirely concrete: just think about it. Consider first the boundary condition for S; = 0. If the stock price at time t is zero, it will be zero forever. To see this, either note that the stochastic differential equation for S; becomes dS; = 0 at time t, and hence the stock price never changes, remaining at zero. Alternatively, recall the solution to the stock Price stochastic differential equation is given by Sp vex {(r— Jo?) -04 own wa},m 2. OPTION PRICING so if 5; is zero then so is Sr. Thus the call option will be worthless, and we have the boundary condition C(¢,0) = 0, t € [0,7]. As a more concrete approach, if we substitute S; = 0 into the Black-Scholes equation, we end up with ac x By 60) = 70.0) This is an ordinary differential equation which we can solve to give C(t,0) = e*C(0, 0). We know C(T,0) = max{0 — K,0} = 0. This gives C(0,0) = 0, and in turn this implies C(t,0) = 0 for all t. The boundary condition at S; = oo is a little harder to specify. For very large values of S;, the option is almost certain to finish in-the-money. Thus for every dollar the stock price rises, we can be almost certain to receive a dollar at payoff, time T. This is sometimes written as i sot00 OS t, 8) = 1 Alternatively, one can observe that as the option gets deeper and deeper into the money, the optionality gets worth less and less so the boundary condition is that C=3,-K for S; large. Note this is only one way to derive the Black-Scholes equation and it is wise to know many ways. For further details on the Black-Scholes equation and related background, see [6]. Here are some possible related questions: « If the payoff function is instead FS) for some deterministic function F, what are the boundary conditions at S; = 0 and 5; = 00? ‘* Prove that the equation S; = So exp {(r — 07) t + oW;} satisfies the stochastic differential equation given for S.. ‘© Prove that the Black-Scholes formula for a European call option satisfies the Black-Scholes equation23. SOLUTIONS 25 Derive the equation if the stock pays continuous dividends at a rate d. ¢ Transform equation (2.1) into the heat equation using a change of variables. a Solution to Question 2.2. What sort of undergrad are we dealing with here? Obviously there is a large difference between a student directly out of high school and one nearing the end of their studies in probability theory or financial mathematics. The best interpretation of this question is to give an explanation which is as simple as possible. ‘One unavoidable, and somewhat technical, statement is that in the Black- Scholes world the arbitrage-free stock price evolves according to the stochastic differential equation dS, = rSidt + 05d, where r is the risk-free rate of return (whether the undergrad understands much stochastic calculus is questionable, but short of giving a brief explanation of what the above equation represents there is little we can do to avoid using this). Here you should mention that ‘arbitrage-free’ essentially implies that there does not exist opportunities to make money for nothing without any risk in the market. One could also give an elementary explanation of what this equation represents; see the extension questions below. We require one other asset to use in the derivation, the risk-free bank account. This grows at the continuously compounding rate r and hence its value at time t, By, is given by By =e > dB, = rBedt, which is a result from ordinary calculus. ‘The final necessary piece of technical mathematics we require is It6’s formula: the stochastic differential equation of a function f(t, St) is given by of 18°F df(t, oral 5 H as 208? Evaluating this requires the relations (dt)? = (dW;)(dt) = 0, (dW,)? = dt. Here we can compare this result to those from ordinary calculus, noting the extra a, sydes (8)(a8,)? (4, Si)dSe + term as a consequence of differentiation using stochastic processes.