Deutche Bank

9-205-059 REV APRIL 26, 2005 GEORGE CHACKO PETER HECHT VINCENT DESSAIN ANDERS SJOMAN Deutsche shore conclusion Relative- pass judgment Trades It was the third week of fantastic 2003, and Jamil Baz, head of Deutsche coin banks furbish up Income search Group, gathered his query base for a morning meeting. So, what ar the markets telling us now? he asked the group. Are there some(prenominal) trends or freshs for new c atomic number 18 ideas? The contumacious Income search Group that Baz led was Deutsche depository financial institutions internal question and development (R&D) department for unflinching income promoters.Their mandate was to tonus for untapped value across seize markets and interest prise derivatives. Long-term-oriented investigate findings were presented to clients, whereas immediate opportunities were suggested as guiles to internal traders as well as clients. The victory of the group was in part measured by how many of their trade suggestion s actually turned into achievementful trades. So far, they had achieved an impressive 75% success rate. A natural place to start vistaing for new trades was the up-to-the-minute prices on various U.S. exchequer bonds (see Exhibit 1 for data from exalted 15, 2003). The groups members consistently went through that data set, looking for come-at-able trades to recommend. Typically relative-value trades took both(prenominal) pertinacious and bunco positions across divergent separate of the succumb slide. Bazs standard weekly question just emphasized what they all knew that it was time to scour through the numbers one and only(a) more(prenominal) time to see if any such positions were available. The Deutsche Bank Fixed Income Research Group notchquartered in Deutsche Banks London office, the companys Fixed Income Research Group consisted of astir(predicate) 50 analysts and strategists. (An additional 10 were regain in the buzzwords New York offices. ) Global head of Fixe d Income Research and in charge of the group was Baz, a managing director with Deutsche Bank since 2001. Previously at Lehman Br otherwises in London, Baz also held an M. S. in management from MIT and a Ph. D. in business economics from Harvard University. As a part of a large financial institution, the research group was under constant blackjack to monetize the ideas that they generated.The group presented its findings both internally to the ________________________________________________________________________________________________________________ Professors George Chacko and Peter Hecht, Executive handler of the HBS Europe Research Center Vincent Dessain, and Research Associate Anders Sjoman prepared this incase. This case deals with trade-specific advice activities of a research department and draws heavily from Deutsche Bank Discussing the Equity assay Premium, HBS Case No. 205-040, by the comparable authors. Case No. 205-040 deals with macro-level advice from the res embling research department.Some names and data have been disguised for confidentiality. HBS cases are essential solely as the basis for class discussion. Cases are not think to serve as endorsements, sources of primary data, or illustrations of effective or otiose management. This case is not intended as financial advice, and it should not be used as the basis for any investment decision, in unscathed or in part. Copyright 2004 President and Fellows of Harvard College. To order copies or collect permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http//www. bsp. harvard. edu. No part of this publication may be reproduced, stored in a retrieval system, used in a bedcover winding-sheet, or transmitted in any form or by any blind drunkselectronic, mechanical, photocopying, recording, or otherwisewithout the permission of Harvard Business School. 205-059 Deutsche Bank Finding Relative-Value Trades Deutsche Bank traders, as well as externally to Deutsche Bank clients at the CEO, CFO, and treasury level. Baz explained how the ideas were pitched The final goal is to create a right with primed(p) income clients.So, for clients on the asset side, such as mutual funds, hedge funds, insurance companies, and pension plans, we help them generate high takes on their assets. We experience specific ideas to be executed by the clientshopefully with us, although that is never certain. However, withal if we dont get a trade out of our recommendation, it is big enough that we maintain Deutsche Banks presence at the client. sometimes we also do bespokeor customizedwork, where we analyze their balance sheet and asset-liability mismatches for them, almost like technical financial consulting.In general, research alone allow not give us clients, but research combined with price are the keys to building long-lasting relationships with external clients. Overall, we strive to push the frontiers of analytic finance when it comes to lessoning interest rates, volatilities, and spreads. Owing to data availability and an inner exposure to institutional market realities, we are often pushed to reach results forrad of academic finance journals. On a group level, Deutsche Bank organise its fix income activities in the global markets around three main pillars investor coverage, issuer coverage, and research.The trading desks dealing in these areas were in turn divided into two groups acknowledgement (with credit trading/credit derivatives, new issue syndicate, asset securitization, and uphill markets) and rates (with foreign exchange, gold markets, fixed income, and interest rate derivatives). The research efforts of the group were set up to match these organizational divisions. The Fixed Income Research Group was one of several research groups (as sh feature in Exhibit 2). only these groups were run under the banner of Global Markets Research.Research as a undivided was he aded by David Folkerts-Landau. Demand for direct meetings with Deutsche Banks research groups had adult over the past few days, taken internally as a sign of increased respect for the banks research takings. In the last year, Bazs group alone had logged over 1,500 client meetings. tout ensemble clients had access to the Deutsche Bank research in text file and newssheets that were available online. midland traders also benefited from the research, which was a major influence behind practically of the banks patentedor proptrading.Most members of the research group overlap their time between external clients and traders, with more senior staff members functional more with external clients and less with the trading floor. In the end, measuring the research groups value to the organization was still difficult. Said Baz pose a value on the work we do, and the effect we have on the bank, is very hard. In fact, if you were to really measure it by attributing sales and trades pu nt to us, the trading floor would be more reluctant to work with us. Instead, we are mostly evaluated by top management on three other factors.Firstly, overall market direction, which is how much of rate and spread moves did we catch in our advice. Secondly, the relative-value trades we originated. Thirdly, any customized business we have brought in from our client meetings. Compensation to the members of the group was tied to the evaluation of the group as a whole. Individual bonuses were past given at the discretion of Baz as the groups manager, found on his qualitative impression of each members contribution. 2 Deutsche Bank Finding Relative-Value Trades 205-059 Strategic Advice and Relative-Value ActivitiesThe groups activities were normally upset down into strategic advice on macro trends and relative value. The strategic advice activities streng consequentlyed on long-term discussions with clients, where the group presented Deutsche Banks view on macroeconomic trends to ex ternal clients. In these discussions, George Cooper, the groups global fixed income strategist, typically did not expect a quick monetary return. Cooper, a Ph. D. receive in engineering at Durham University with experience from both Goldman Sachs and JP Morgan, explained This type of activity does not generate a lot of money from a trade perspective.It generates brand value, though, and is especially appealing to insurance companies or asset-liability people, who appreciate the long-term view. We believe it serves more of an educational purpose. It gets the fund managers thinking. They are not looking for prescriptive research, where we tell them to do this trade, but they look for interesting ideas. Of course, they then weigh our ideas against whatever Goldman Sachs or Morgan Stanley are saying. Our usage is to come up with hopefully insightful but also enlightening new ways to look at things.By contrast, the relative-value activities looked for more immediate opportunities by c omparing different instruments and then recommending various trading strategies to clients and internal traders. Head of Relative Value Research for Europe was Jean Dumas, an engineer from ESME SUDRIA in France with a specialization in finance, who had worked with Relative Value Research for Deutsche Bank in Paris, Frankfurt, and Sydney before moving to London. Dumas explained his work We come up with different types of trades all the time.The trade opportunities may be there for a week or two, sometimes longer. I look at different spreadsheets, find out to what traders are saying, watch the news, study different influences. . . . Then I try out to put everything togetherand suddenly there is a trade opportunity. Our telephone line is really grabbing things that dont seem to be related at first and see if there is a trade to be done. The trade opportunities that the research group identified were published weekly in the newsletter Deutsche Bank Fixed Income Weekly, which was dis tributed to Deutsche Bank traders as well as to clients.A sponsor contributor to the newsletter was Dr. Nikan B. Firoozye, head of Global decimal Strategies and a Ph. D. graduate in mathematics from Courant Institute at New York University with experience from adhesiveness Capital, Sanford C. Bernstein, and Lehman Brothers. Firoozye explained I write a piece on Euroland strategy every week where we suggest trades. Some of these are big trades that we dont change very often, such as curve-steepening trades. We can have the same trade off and on for a full year. We also add up economic data as it impacts the bond markets.For instance, how structured trades could be influenced by the move in dollar versus yen, and how you should position for that. In his role as head of Euroland Strategy, Firoozye also oversaw all strategic investments in Euroland bond markets. He was also involved in all modeling issues and wrote stand-alone papers on quantitative strategy. Looking for a Relative-V alue Trade For the research group, one way to find relative-value trades was to compare the prices of traded securities against the prices that the group thought the securities should trade at. This inwrought view 205-059 Deutsche Bank Finding Relative-Value Trades was based on a proprietary model developed at Deutsche Bank. (Most banks used proprietary models as a base from which to evaluate the prices of traded securities. ) The models were built on the fact that the returns offered by fixed income instruments could be characterized by the yields that they offered. The yield was roughly seen as payment for the risk of exposure borne by the holder of that security. There were many sources of risk in fixed income securities, such as interest rate risk, credit risk, and prepayment risk.Also, the yield of an instrument could be broken down into components. The components could be thought of as requital for the different sources of risk. So, for example, the yield on a corporate bo nd could be thought of as being composed of a risk-free yield plus a credit spread. The risk-free yield represented compensation for interest rate risk in the bond, while the credit spread represented compensation for default risk in the bond. 1 To guess the compensation for the interest rate risk alone, banks typically constructed yield-curve models. These were models for the yields on zero-coupon treasury securities, since Treasury instruments typically contained only interest rate risk. Models for the yield curve could be then used to compare the current and expected prices of U. S. Treasury instruments. 2 The research group at Deutsche Bank had developed their own proprietary yield-curve model, a so-called three-factor affine model (see Exhibit 3 for a conceptual description of the model). Firoozye explained the fundaments of the model We have three factors private road the yield curve that we see as analogous to the economy.In an economy, there is inflation, output gaps, and short rates. So first among our factors is a long rate, which is analogous to inflation. It is the long-playing mean reverting of our three factors. In the fifties inflation was low, in the seventies it was extremely high, and now it is back down again. It takes 20 years to go through its cycle. It is very slow, very persistent, whereas the business cycle is much, much faster. You go through a business cycle in about seven years. So slope, our second factor, is then the measure of output gap. run mean reverts much more quickly than inflation.The third factor is the short rate, which mean reverts the fastest. later on estimating the variables of the three-factor model, the team calibrated the model to price the one-month, two-year, and 10-year zero-coupon bond. After Bazs request at the weekly meeting, the analysts now used the latest numbers on various U. S. Treasury bonds to update and calibrate the model (see Exhibit 4 for the resulting output from Deutsche Banks model). The i dea was to then compare the actual zero-coupon yield curve against the predicted ones coming out of the model and see if any trade ideas presented themselves.In fact, several trades seemed to come out of that comparison. Baz and the team now had to pick the trades with the highest profit potential. 1 It should be noted that the fantasy of compensation here is approximate. The yield on a zero-coupon corporate bond is not the expected return of that bond. It is simply the promised return of that bond, or the return an investor would get if the bond did not default. Starting with this promised return and then factorisation in the probability of default and a default risk pension leads to the expected return for that bond. More generally, yield-curve models could be used to price any interest rate-sensitive security. For example, the pricing of interest rate options starts with a yield-curve model. 4 Deutsche Bank Finding Relative-Value Trades 205-059 Exhibit 1 Prices and Coupon Rate s of Various U. S. Treasury Bonds on August 15, 2003 Coupon Rate (%) 3 2. one hundred twenty-five 1. 5 6. 5 5. 625 2. 375 6. 25 3. 25 3 3. 25 5. 5 6 6. 5 5. 75 5 5 4. 875 4. 375 3. 875 4. 25 13. 25 12. 5 11. 25 10. 625 9. 25 7. 5 8. 75 8. 875 9. 125 9 8. 875 8. 125 8. 5 8. 75 7. 875 8. 25 8 7. 25 7. 125 6. 25 7. 5 7. 5 7. 625 6. 875 6 6. 75 6. 625 6. 375 6. 125 5. 5 5. 25 6. 125 Maturity Date 2/15/2004 8/15/2004 2/15/2005 8/15/2005 2/15/2006 8/15/2006 2/15/2007 8/15/2007 2/15/2008 8/15/2008 2/15/2009 8/15/2009 2/15/2010 8/15/2010 2/15/2011 8/15/2011 2/15/2012 8/15/2012 2/15/2013 8/15/2013 2/15/2014 8/15/2014 2/15/2015 8/15/2015 2/15/2016 8/15/2016 2/15/2017 8/15/2017 2/15/2018 8/15/2018 2/15/2019 8/15/2019 2/15/2020 8/15/2020 2/15/2021 8/15/2021 2/15/2022 8/15/2022 2/15/2023 8/15/2023 2/15/2024 8/15/2024 2/15/2025 /15/2025 2/15/2026 8/15/2026 2/15/2027 8/15/2027 2/15/2028 8/15/2028 2/15/2029 8/15/2029 Current Price 101. 0544 100. 9254 99. 8942 109. 0934 108. 438 99. 7848 111. 7184 101. 0841 99. 1692 99. 271 109. 7707 112. 145 114. 9084 110. 3894 105. 2934 104. 7607 103. 4391 99. 2806 95. 0288 97. 7693 174. 3251 168. 9389 157. 0552 152. 4222 140. 0135 123. 3044 136. 0598 137. 504 140. 792 139. 9079 138. 7431 130. 7162 135. 2938 138. 3466 128. 4995 131. 7341 130. 4736 121. 58 120. 1744 109. 4538 125. 46 125. 4466 127. 1477 117. 5509 106. 3626 116. 1986 114. 7086 111. 036 108. 0391 99. 633 96. 2876 108. 4062 outset Adapted by casewriter from Datastream. 5 205-059 Deutsche Bank Finding Relative-Value Trades Exhibit 2 Deutsche Bank Global Markets Research Organization Global Head murmur and Co-Head Global economics Global Strategist Foreign Exchange Securitization mogul Development Regional Heads Asia/ Pacific Germany Strategy Fixed Income/ Relative Value Research Emerging Markets Credit Credit (High Grade Credit Research) Economics Global US Europe Emerging Markets Source Deutsche Bank. Exhibit 3 Deutsche Banks Zero-Coupon Yield Model Key variables Short rate, slope, and long rate (or short rate, output gap, and inflation) Model specified by a system of equations (in Q measure) Long rate mean reverts slowly (possibly to nonzero mean) dX t= ( X ? k X X t) dt + ? X dWt X Slope mean reverts faster (to zero) dYt = ? kY Yt dt + ? Y dWt Y X t + Yt ? rt = 0 In equilibrium short rate, rt, follows the target Xt+Yt (an analogue of the Taylor rule) Short rate mean reverts fast in order to restore the equilibrium drt = k r ( X t + Yt ? rt ) dt + ? r dWt rSource Adapted by casewriter from Quantitative Models for Fixed Income, Deutsche Bank presentation, October 2003. 6 Deutsche Bank Finding Relative-Value Trades 205-059 Exhibit 4 Output from Deutsche Banks Zero-Coupon Yield Model Model prognostication (BEY) 1. 2443% 1. 8727% 2. 4110% 2. 9665% 3. 4454% 3. 8557% 4. 1996% 4. 4677% 4. 6528% 4. 7107% 5. 7160% 5. 9517% 5. 9315% Maturity (years) 1y 2y 3y 4y 5y 6y 7y 8y 9y 10y 15y 20y 25y Source Note Adapted by casewriters from Deutsche Bank information. The yields in this table are bond equivalent yields (BEY), that is, the semiannual yield reckon by two. 7