Concepts, Components and Collections of Trading Strategies and Market Color
aa r X i v : . [ q -f i n . GN ] J a n Concepts, Components and Collections of Trading Strategies andMarket Color
Ravi KashyapSolBridge International School of Business / City University of Hong Kong
January 14, 2020Keywords: Trading Strategy; Investment Hypothesis; Uncertainty; Trial and Error; Risk Management;VolatilityJEL Codes: G11 Investment Decisions; D81 Criteria for Decision-Making under Risk and Uncertainty; C63Computational Techniques
Contents Abstract
This paper acts as a collection of various trading strategies and useful pieces of market information that mighthelp to implement such strategies. This list is meant to be comprehensive (though by no means exhaustive)and hence we only provide pointers and give further sources to explore each strategy further. To set thestage for this exploration, we consider the factors that determine good and bad trades, the notions of marketefficiency, the real prospect amidst the seemingly high expectations of homogeneous expectations from humanbeings and the catch-22 connotations that arise while comprehending the true meaning of rational investing.We can broadly classify trading ideas and client market color material into Delta-One and Derivative strategiessince this acts as a natural categorization that depends on the expertise of the various trading desks that willimplement these strategies. For each strategy, we will have a core idea and we will present different flavors ofthis central theme to demonstrate that we can easily cater to the varying risk appetites, regional preferences,asset management styles, investment philosophies, liability constraints, investment horizons, notional tradingsize, trading frequency and other preferences of different market participants.
We can broadly classify trading ideas or strategies (Pardo 2011; Nekrasov 2014; End-note 1) and marketcolor (Merkley, Michaely & Pacelli 2017; Cheng, Liu & Qian 2006; Groysberg, Healy & Chapman 2008;Schipper 1991; Williams, Moyes & Park 1996; End-note 2, 3, 4) material into Delta-One and Derivativestrategies (generally, the price of Delta one securities have a one to one correspondence with the price of theunderlying: Natenberg 1994; Hull & Basu 2016; End-notes 5, 6), since this acts as a natural categorizationthat depends on the expertise of the various trading desks that will implement these strategies. For eachstrategy, we will have a core idea and it is very straight forward to present different flavors of this centraltheme that can easily cater to the varying risk appetites, regional preferences, asset management styles,investment philosophies, liability constraints, investment horizons, notional trading size, trading frequencyand other preferences of different market participants. Once we have some market color or a possible tradingidea, the conundrum seen in the title of this section presents itself, which is whether to implement it andhow exactly; we address some of these concerns through the rest of this article.As the number of securities in any trading idea is varied, the error in the expected returns, the variance ofreturns, the benefits of diversification and the liquidity constraints will vary (risk, return, diversification andliquidity are discussed in: Bodie, Kane & Marcus 2014; Elton, Gruber, Brown & Goetzmann 2009). Hence,it is important to consider different number of securities as different variations of the core strategy. Thecosts of implementing the strategy will also vary with the number of securities and other parameters used tocreate the strategy. Hence, it is important to explicitly calculate these figures for different variations of thecore strategy (for trading costs see: Perold 1998; Almgren & Neil 2001; Kashyap 2016b).2he Execution Risk and the Post Execution Portfolio risk of the strategy will depend on the Market Impact,Market Risk and the Timing of the executions that make up the strategy (Kashyap 2016b). Ways to achieveoptimality in this regard are elaborated below in the “Risk and Best Execution Advisory” point 10. Anyinvestment firm that already has some of these trading strategies, can look for supplementary informationto improve the pre and post trading experience. Combination strategies (or cross-trading) is possible whenimplementing two strategies reduces the overall market exposure and hence the risk of the blended strategies.Also, in certain cases, once we implement one strategy, the additional cost and the risk to implement anotherstrategy, (marginal cost and risk) is significantly less and hence it is can be demonstrated that it is beneficialto implement multiple strategies.These ideas can be implemented mostly with data available from Bloomberg (Bloomberg 1981; End-note 7)or other market data vendors. If additional information is required it is mentioned along with the specifictrading idea (we might be able to supplement any basic data set with specialized data sources). A statisticaltool like MATLAB (Window 2010; End-notes 8, 9; a basic database to maintain the time-series data andother information would be a prudent investment as well) would be required to perform the regressions andother computations to back-test (Bailey, Borwein, de Prado & Zhu 2014) the strategies, estimate the riskand Profit and Loss (P&L) levels on these strategies. For the sake of brevity and to keep this documentsimply as an overview of the possibilities, the computational complexities and implementation details havebeen abstracted away from the individual strategy sections below. As it will become clear, the design andimplementation of these strategies will require personnel with quantitative skills and training in mathematicsas it applies to finance and economics. In (Kashyap 2018a) we provide a complete numerical illustration ofone trading strategy that can be beneficial when we expect the market to rebound after a slump or after abad economic cycle.This paper acts as a collection of various trading strategies and useful pieces of market information that mighthelp to implement such strategies. This list is meant to be comprehensive (though by no means exhaustive)and hence we only provide pointers and give additional sources to explore each strategy further. To set thestage for this exploration, we consider the factors that determine good and bad trades, the notions of marketefficiency, the real prospect amidst the seemingly high expectations of homogeneous expectations from humanbeings and the catch-22 connotations that arise while comprehending the true meaning of rational investing.
We can represent the investment process as a dotted circle since there is a lot of ambiguity in the varioussteps involved. The circle also indicates the repetitive nature of many steps that are continuously carried outwhile investing. If we consider the entire investment management procedure as being akin to connecting thedots of a circle, then the Circle of Investment can be represented as a dotted circle with many dots falling3pproximately on the circumference, but it hard to have an exact clue about the exact location of the centeror the length of the radius.The Equity asset class holds the potential for unlimited upside and brings with it partial ownership of thefirm and hence some influence over the decision making process. It can be argued that this premise ofboundless profits, coupled with limited losses or liability and a certain degree of control, make this asset classan extremely appealing one, contributing to its immense popularity. Hence, the strategies discussed below aremore immediately applicable to the stocks markets, but it is fairly straight forward to extend them to otherasset classes (for foreign exchange and fixed income securities, see: Copeland 2008; Tuckman & Serrat 2011).The various asset classes can be compared to balloons tethered to the ground, with the equity balloon havingthe weakest connections to the ground and also the weakest controls to guide it, if it is wind-borne. The lackof a strong controlling factor also makes regime changes much harder to detect for equities (regime changesare a major shift in the investment landscape: Wade 2008; Angeletos, Hellwig & Pavan 2007; Gasiorowski1995).We need to keep in mind that good traders can make bad trades and bad traders can make good trades, butover many iterations, the good traders end up making a greater number of good trades and hence we providesome distinctions between good and bad trades below (Kashyap 2014b).1. The factors that dictate a good trade or a bad trade depend on the Time Horizon and the InvestmentObjective. The time horizon can be classified into short term, medium term and long term. Theinvestment objective can be conservative or aggressive. While there are no strict boundaries betweenthese categories, such a classification helps us with the analysis and better identification of trades.2. Any trade that fulfills the investment objective and time horizon for which it is made is a good trade.Otherwise, it is a bad trade.3. On the face of it, we can view good trades as the profitable ones and bad trades as ones that losemoney. But where possible, if we try and distinguish between proximate causes and ultimate reasons,it becomes apparent that good trades can lose money and bad trades can end up making money.4. Due to the nature of uncertainty in the social sciences: the noise around the expected performance ofany security; our ignorance of the true equilibrium; the behavior of other participants; risk constraintslike liquidity, concentration, unfavorable Geo-political events; etc. implies we would have deviationsfrom our intended results. The larger the deviation from the intended results, the worse our trade is.5. What the above implies is that, bad trades show the deficiencies in our planning (estimation process)and how we have not been able to take into account factors that can lead our results astray. It is truethat due to the extreme complexity of the financial markets, the unexpected ends up happening and4e can never take into account everything. We just need to make sure that the unexpected, even if itdoes happen, is contained in the harm it can cause. The good thing about bad trades is the extremelyvaluable lessons they hold for us, which takes us through the loop or trials, errors and improvements.6. We then need to consider, how a good trade can lose money. When we make a trade, if we knowthe extent to which we can lose, when this loss can occur and that situation ends up happening, ourplanning did reveal the possibility and extent of the loss, hence it is a good trade.7. The bottom line is that, good trades, or bad trades, are the result of our ability to come up withpossible scenarios and how likely we think they will happen. The ability to visualize possible outcomesis related to intelligence. Everything else being equal, someone with experience or someone who has hadthe benefit of having participated in repetitions of similar situations and then having made decisions insome cases after mistakes in earlier iterations (learning through trial and error: Ismail 2014; Kashyap2017; End-note 10) will be better at facing uncertain situations and solving problems.Zooming back into equities, the following are some other factors that can contribute to good equity trades.1. The trade will not soak too much of the available liquidity, as measured by the average trading volume,unless of course, we wish to take a controlling stake in the firm (Bernstein 1987; Pástor & Stambaugh2003; Bhide 1993; Amihud 2002; Hameed, Kang & Viswanathan 2010).2. It is held by a number of investors. There is more uncertainty if there are more investors, but it seemsto work to our benefit in most cases. If the number of investors is limited, the possibility of all ofthem doing the opposite of what we want is higher and more likely (Amihud, Mendelson & Uno 1999;Lakonishok, Shleifer & Vishny 1992; Chan & Lakonishok 1993; Gompers & Metrick 2001; Asquith,Pathak & Ritter 2005; An & Zhang 2013).3. The noise or the randomness is less so that our decisions can be more accurate. This can be measuredby volatility or the price fluctuations that we see (French, Schwert & Stambaugh 1987; Schwert 1989;Ang, Hodrick, Xing & Zhang 2006; Glosten, Jagannathan & Runkle 1993).4. The firm issuing the securities is not too dependent on any particular product, profits from a particularregion, is not overburdened with debt, is paying dividends consistently, its price is not too high comparedto its earnings and other fundamental research indicators (Lancaster 1990; Spence 1976; Bilbiie, Ghironi& Melitz 2012; Randall & Ulrich 2001; Srinivasan, Pauwels, Silva-Risso & Hanssens 2009).5. If we are able to see some pattern in the share price changes, that is a good trade. This means that thissecurity is exhibiting Non-Markovian behavior (Turner, Startz & Nelson 1989; Hassan & Nath 2005;5ai 1994; Hassan 2009; Litterman 1983). Such behavior is usually hard to detect, but it comes downto the lens we are using to view the world or the methods we are using to perform historical analysis(Hamilton 1994; Gujarati 2009; Verbeek 2008).6. If the security is affected by any asset price bubbles and we are able to detect the formation of suchbubbles (Siegel 2003; Stöckl, Huber & Kirchler 2010; Hott 2009; Andersen & Sornette 2004; Scherbina& Schlusche 2014; Kashyap 2016a).7. If we are shorting the security and it has a greater tendency for a downward movement, as exhibited byits skew and other higher moments (Corrado & Su 1996; Bakshi, Kapadia & Madan 2003; Badrinath& Chatterjee 1991; Badrinath & Chatterjee 1988; Chen, Hong & Stein 2001; Barberis, Mukherjee &Wang 2016; Amaya, Christoffersen, Jacobs & Vasquez 2015).8. So far, we have talked about the unknowns that we know about. What about the unknowns that wedon’t know about. The only thing, we know about these unknown unknowns are that, there must bea lot of them, hence the need for us to be eternally vigilant (Taleb 2007).
The analogy of, building a plane and flying it, to constructing a model and trading with it, will help usconsider the associated risks in a better way. Modeling would be the phase when we are building a plane,and the outcome of this process is the plane or the model which we have built; trading would then be theact of flying the plane in the turbulent skies, which are the financial markets. The modelers would thenbe the scientists (also engineers) and the pilots would be traders. It is somewhat out of the scope of thisdocument to discuss questions regarding what kind of person can be good at both modeling and trading.Trading would need a good understanding of what the model can do and where it will fall short; and buildinga model will need to know the conditions under which an model has to operate and the sudden changes thetrading environment might encounter.A deterministic world can be made to seem stochastic quite easily, since randomness is only from the pointof the viewer, the creator of uncertainty (also perhaps, the universe) has no randomness. Hence, it mightbe possible to start with a few simple rules that makes sense intuitively and explain the stochastic behaviorof most phenomenon. Our investment decisions are made over time and so we set the direction of forwardmovement in time to be equivalent to flying the plane forward. Since we cannot see what will happen in thefuture; to fly the plane forward, we should not be able to see what is in front of us. This is equivalent to aplane with the front windows blackened out. All we have are rear view mirrors (most planes don’t exactlyhave rear view mirrors, but let us imagine our plane having one) and windows to the side.6s we are cruising along in time, what we have with us is the historical data or the view from behind andreal time data which is the view from the side, to aid in navigating our way forward or to the future. We usethe historical data to build our model and then use the data from the present to help us make forecasts forwhat the future holds. The modeling would involve using data inputs to come up with outputs that can helpus decide which securities to pick, or to help set the direction of motion. The trading aspect would involveusing the model outputs and checking if that is the direction in which we want to be heading, that is actuallydeciding which securities to pick, and watching out for cases where the predictions are not that reliable.We can use multi-factor models (Hamilton 1994; Ng, Engle & Rothschild 1992) to decompose overall portfoliorisk and help identify the important sources of risk in the portfolio and link those sources with aspirationsfor active return. We need to use the right principles, the right material and the right processes.1. The right principles would require understanding certain concepts that determine the relevant measureof risk for any asset and the relationship between expected return and risk when markets are tendingtowards equilibrium. Examples for these are the Capital Asset Pricing Model (CAPM), the ArbitragePricing Theory or other multi index models (Sharpe 1964; Ross 1976; Roll & Ross 1980; Bodie, Kane& Marcus 2014).2. The right material translates to having data on the security returns and choosing the relevant factors.The amount of data and factors that is available is humongous. We need to use some judgment regardinghow much history to use. We also need to be attuned to significance and causality among the factors.All this can involve some independent data analysis (Krejcie & Morgan 1970; Granger 1981; Adcock1997; Hair, Black, Babin, Anderson & Tatham 1998; MacCallum, Widaman, Zhang & Hong 1999;Lenth 2001).3. The right process would mean using judicious concepts from econometric / statistical theory (Bishop,Fienberg & Holland 2007; Eisenbeis 1977). Some examples would be to check for the stationarity ofvariables, to normalize the variables to scale them properly, to see if there is any correlation betweenthe independent variables and correcting for it (Multi-Collinearity: Hamilton 1994; Maddala & Lahiri1992; End-note 11). We need to make sure no variables that would have an impact are left out (OmittedVariable Bias: Hamilton 1994; End-note 12).There needs to be a lot of tinkering; this means we need to have a continuous cycle of coming up with aprototype, testing how it works and making improvements based on the performance. This is especiallyimportant in the financial markets, since we are chasing moving targets as the markets have a tendency tobe quasi-equilibriums (we never know what a true equilibrium is but perhaps, the markets fluctuate between7ultiple equilibriums, somewhat like a see-saw: Mantzicopoulos & Patrick 2010; Stocker 1998; End-note 13).Modeling needs to be well thought out, with due regard to anticipating as many scenarios as possible andbuilding in the relevant corrective or abortive mechanisms when adverse situations occur. Given that, we arenever close to accomplishing a perfect model, which can handle all cases without failure and without constantchanges, we would need to constantly supervise the outcomes; hence models that are simple and robust arebetter suited, since it is easier to isolate the points of failure when things get rough. Robust here meansproducing similar results under a variety of conditions, with some changes to the inputs or the controls.
A primary question that arises in finance is: whether markets are efficient? Questions & Answers (Q&A)are important. But Definitions & Assumptions (D&A) are perhaps, more important. The good news is that,Q&A and D&A might be in our very DNA, the biological one (Alberts, Johnson, Lewis, Raff, Roberts &Walter 2002; End-note 14). Maybe, DNA hold the lessons from the lives of every ancestor we have everhad. Evolution is constantly coding the information, compressing it and passing forward, what is needed tosurvive better and to thrive, building what is essential right into our genes (Church, Gao & Kosuri 2012;Lutz, Ouchi, Liu & Sawamoto 2013; Kosuri & Church 2014; Roy, Meszynska, Laure, Charles, Verchin & Lutz2015).The assumption made in finance regarding homogeneous expectations (Levy & Levy 1996; Chiarella & He2001), especially in the derivation of the efficient frontier, the CAPM and the Capital Market Line (Bodie,Kane & Marcus 2014), is stunningly sophisticated, yet seemingly simplistic. Most people would argue thatno two people are alike, so this assumption does not seem validated (Valsiner 2007; Buss 1985; Plomin &Daniels 1987). Then again, this assumption is perhaps a very futuristic one, where we are picking the besthabits and characteristics, from our fellow beings (maybe not just humans?) and at some point in the future,we might tend to have more in common with each other, fulfilling this great assumption, which seems moreof a prophecy. Again, if we become too similar then mother nature, or, evolution, will have less to work with;since more differences tell her which characteristics are better for certain conditions and many possibilitiescreate stronger survival potential (Rosenberg 1997; Wilke, ... , & Adami 2001; Elena & Lenski 2003; Nei2013). Too much similarity might not be an issue if survival is no longer a concern (Kashyap 2018b). Butwith respect to finance, we might evolve enough, so that one day, we might have the same expectations withrespect to our monetary concerns. This would also be the day when the Bid-Offer spread would cease tomatter, or, we would be indifferent to it, making every coffee shop, theater, street corner, pub, or everywhere. . . a venue for any product (Kashyap 2015).Hence if we vary our definitions of efficiency and the corresponding assumptions, a useful chain of thoughts8nd efforts would be to capture, whether and how much, the actions of various players takes the marketstowards different forms of efficiency (Fama 1970). This implies a belief that different markets could havedifferent levels and forms of efficiencies over different times and understanding how the players are actingcould be useful to predict what information could be an advantage, depending on the efficiency that is believedto be at work.Perhaps, another way of looking at and understanding this concept is by pondering over the notion of arational investor, who has been defined in multiple ways and extensively discussed (also known in economiccircles as Homo Economicus: Persky 1995; Thaler 2000; End-note 15). Before we consider the Q&A / D&Arelated to a rational investor, we state a simple game from game theory called “Guess 2/3 of the average”(Nagel 1995; another related game is known as the Keynesian beauty contest: Keynes 2018; Büren, Frank &Nagel 2012; Nagel, Bühren & Frank 2017; End-note 16). The objective of the game is for the participantsto guess what 2/3 of the average of the guesses of everyone participating in the game will be, and where thenumbers are restricted to the real numbers between 0 and 100, inclusive. The winner is the person whosechosen number is closest to the 2/3 average of all chosen numbers and will obtain a fixed amount as a payoffthat is independent of the stated number and 2/3. If there is a tie, the prize is divided equally among all thewinners.There is a unique Nash equilibrium (Nash 1951; Osborne & Rubinstein 1994; End-note 17) for this gamethat can be found by iterated elimination of weakly dominated strategies. Suppose that all participants guessthe highest possible number, 100, then clearly the average of all the guesses will be 100. But to win, theyneed to guess 2/3 of the average, hence their guess has to be 2/3*100, but if everybody has 2/3*100 as theirguess, they need to guess 2/3 of 2/3*100 or 2/3*2/3*100 and this process will continue. Continuing thisline of reasoning k times and as k gets larger and larger, that is as k → ∞ , the guess g of any participantgets closer and closer to zero. Alternately stated, the limiting value for the guess game, g , as the numberof iterations k increases, (as we continue our line of reasoning by taking the / of the maximum value thatsomeone can guess) is given by, g = lim k →∞ (cid:18) (cid:19) k ∗
100 = 0 (1)This game is usually played to demonstrate that few people, including students of economics and gametheory, actually get the equilibrium answer of zero (End-note 18). Hence if a rational participant is someonewho would want to win this game by making the right decisions and obtain the winning payoff, it is unclearwhether he should present the above equilibrium guess (Eq 1). This example illustrates that other thanthe non-trivial mathematics expected from any rational participant, he is supposed to be anticipating whateveryone will do, which could include the possibility that not everyone participating might be fully rational.Hence his winning guess might have to be different from the one shown in (Eq 1), making him do irrationalthings unless he can not only read the minds of everyone but also force his will upon them to be rational.9o be rational he has make sure everyone acts rationally or he has to act irrationally: a Catch-22 situation(Heller 1999).In a similar vein, we can define a rational investor as someone who is not just factoring in the potentialinvestment decisions of every market participant, but is also having to ensure that everyone is making thedecisions that will be beneficial to him. In the above game, the decisions of everyone had to be the same, butit could be different in many other situations, especially in the financial markets (we will not discuss furtherthe very stimulating topic of whether and how two different decisions can still be beneficial and rational tothe participants; but it would perhaps be proper material for a slew of papers and books). The messagefor us is that either we all have to become rational investors or nobody can be a rational investor. Anotherimplication of this chain of thoughts is that if all of us become rational investors then our expectations mightbecome homogeneous, making markets efficient and validating the assumptions made to derive finance andeconomic theories; though if that happens there might be no need for economics and finance at that point,since we might have transcended beyond the need for wealth, wants and other worldly aspects: anotherCatch-22 situation.The computational challenge in this case was limited due to the rather simplified nature of the exercise.But in other decision making situations encountered in daily life, the tools and theories of economics pre-scribe solving complex optimization problems to arrive at the right results. If we ponder on the intellectualrequirements an agent has to possess to confront and navigate his daily challenges, it becomes clear that itis beyond the capabilities of all humans to use such calculations in their everyday decisions. We want tohighlight here that some humans have created complex but wonderfully elegant solutions for many simpleeveryday problems. But even after assuming that such a human would use the solution he has developed fora particular problem, he would fail to adopt the solutions created by others for other problems he encounterssince he would have limited expertise with problems in another domain.Clearly, another case is where we might not have developed a solution ourselves, but we have studiedthe solutions of others extensively enough so that we are intimately familiar with it and using it has becomesecond nature to us. Even if we make the valid argument that many solutions are related and the marginaleffort to master a new solution is less once someone has good knowledge of a few solutions, we need to beaware that there are too many new techniques and tools being introduced and the complexity in solutions isincreasing significantly. This issue of most of us not being familiar with most solutions is exacerbated withthe highly specialized nature of academic research being conducted and encouraged, coumpounded by theartificial boundaries we have created by labeling disciplines. Many journals do state that they encouragemulti-disciplinary work but when confronted with work that truly transcends the fake silos of knowledge wehave created, editors and reviewers struggle to make the right decisions: (Ke, Ferrara, Radicchi & Flammini2015 study how common , “sleeping beauties are in science”, papers whose relevance has not been recognized10or decades, but then suddenly become highly influential and cited; Gans & Shepherd 1994 find that journalshave rejected many papers that later became classics; this should make us aware of the possibility that manypapers are being rejected only because their contributions are not being realized).We want to emphasize that these are unintended consequences due to the constraints placed on theactual channeling of research efforts to knowledge creation and dissemination. One reason why such unwantedoutcomes creep up, despite the fact that journals, editors, scholarly associations and other research institutionsare wonderful innovations done with the honorable intention of helping us comprehend the cosmos aroundus, is because we live in a world that requires around 2000 IQ points to consistently make correct decisions;but the smartest of us has only a fraction of that (Ismail 2014; Kashyap 2017; End-note 10). Hence, we needto rise, above the urge to ridicule, the seemingly obvious blunders of others, since without those marvelousmistakes, the path ahead will not become clearer for us. Someone with 2000 IQ points is surely a superhero, aptly named IQ-Man. So a rational investor is this mythical character called IQ-Man, who unlikeother super-humans like Super-Man does not even have a movie about him (for society’s fascination withsuperheroes or super-humans see: Reynolds 1992).
1. Index Re-balanceWhen an index is re-balanced, certain constituents are removed, added or their weights in the indexare changed. Many participants try to anticipate these changes and take positions depending on whatthey expect to change in any index. Trading strategies and market color can be created indicatingexpected inflow and outflow at a security level based on the expected weight change or dependingon the re-balancing event (Kostovetsky 2003; Bloom, Gouws & Holmes 2000; Aked & Moroz 2015;Chow, Hsu, Kalesnik & Little 2011). Basket Trading Ideas to take advantage of expected inflow andoutflow of Index Funds based on expected weights on the re-balance dates. We can use Index Tracking(Beasley, Meade & Chang 2003; Guastaroba & Speranza 2012; Dose & Cincotti 2005; Stoyan & Kwon2010) and Co-Integration (Alexander 1999; Engle & Clive 1987; Banerjee, Dolado, Galbraith & Hendry1993; Harris 1995; Alexander & Dimitriu 2005a, b) principles to provide better estimates of the basketsand capture a certain level of dollar flows. MSCI (Hau, Massa & Peress 2009; Chakrabarti, Huang,Jayaraman & Lee 2005; End-note 19) indices are ideal set to start with and it is possible to extend itto other indices with minor adjustments depending on the rules used for the re-balancing that coulddiffer among the many index providers. This strategy would need index membership information from11he relevant index providers. Please see Index Tracking and Co-Integration in points 4, 52. Macro Theme BasketsCreation of stock baskets depending on Macro themes (Burstein 1999; Franci, Marschinski & Matassini2001; Chen, Leung & Daouk 2003). Bounce Basket: Securities expected to benefit the most from therebound in the securities markets and the overall economy (Kashyap 2018a). Short Basket: Securitiesthat are expected to show a significant fall due to fundamental weakness and an overall drop in themarket. Securities that perform the best during upward expected Inflationary moves. Securities thatperform the best during upward movement of Oil Prices, Gold Prices or other commodities. Regionalbaskets that can best cater to investors looking for regional exposure. These baskets are identified byperforming a factor analysis (principal component analysis or regressions can also be used: Hamilton1994; End-note 20) of historical security prices and other factor indicators (Ludvigson & Ng 2007;Jolliffe 1986; Shlens 2005; Costello & Osborne 2005). We can use Index Tracking and Co-Integrationprinciples to provide better estimates of the baskets. Please see Index Tracking and Co-Integration inpoints 1, 4, 5. Macroeconomic data-sources would be required for this set of trading ideas.3. Sector Theme BasketsThese set of strategies are aimed at capitalizing on expected sector rotations. Depending on variousmacroeconomic developments and business cycle trends, different sectors are expected to be eitherbullish or bearish. We can create baskets of stocks to benefit most from these expected trends. Itis possible to put a regional spin on each of the baskets below. We can customize this for differentdurations of the trades and different notional amounts. We can use Index Tracking and Co-Integrationprinciples to provide better estimates of the baskets. Please see Index Tracking and Co-Integration inpoints 1, 4, 5. Sector specific data-sources would be required for this set of trading ideas.The following are some of the sector theme baskets. • Gaming (Casino) Basket • Real Estate Basket • Technology Basket • Financials Basket • Utilities Basket 12
Travel and Luxury Hotels Basket4. Index Tracking BasketsWe can customize this based on two characteristics. Create baskets of stocks to track the index withthe least tracking error and a desired number of stocks. Produce a certain amount of tracking error andselect the number of stocks required to achieve this level of tracking error. Please see Index Trackingand Co-Integration references in point 1.5. Co-Integration BasketsPairs Trading Ideas are based on two portfolios (or even individual securities: Miao 2014) of stockswhich have moved together historically, but are now in diverging positions, so we expect them toconverge back. This is easily implemented for pairs of securities in the same sector. Please see IndexTracking and Co-Integration references in point 1.6. Portfolio Risk BasketMany sell side desks take on principal positions (commit their own capital instead of acting just as anintermediary) on baskets of stocks that are opposite to client trades. A basis point quote is providedto the client based on the market impact to enter the positions, to exit the positions, volatility of thesecurity prices, expected duration to enter the position, expected duration to exit the position andtransaction costs. Inventory Management techniques (Lancioni & Howard 1978; Blackstone Jr & Cox1985) will need to be used to manage the overall exposures on the desk that is creating this strategyand this will be incorporated into the basis point pricing of the risk basket.7. Index Arbitrage Trading IdeasArbitrage between the prices of index constituents and the corresponding futures contracts (Chung1991; Dwyer Jr, Locke & Yu 1996; Kumar & Seppi 1994; Nandan, Agrawal & Jindal 2015; Fung, Lau& Tse 2015). We will need to consider the effect of stock commissions and futures commissions onthe profitability of the trades. We can provide early close out options as opposed to close out on theexpiration date. We will need to consider the market impact of the stock transactions on the stockprice when we put on the trade and the effect of the same if we do an early close out.13. Rich / Cheap Analysis based on Stock BetaWe can create baskets of under-priced and overpriced securities based on the Beta of individual securitiesversus different indices acting as a proxy for the market (Black 1992; Isakov 1999; Amihud & Mendelson1989; Fletcher 2000).9. Equity Swap BasketsIn certain markets, most investors do not have access to trade the securities and the only way to getaccess to these markets is through swaps from a registered broker (Kijima & Muromachi 2001; Gay,Venkateswaran, Kolb & Overdahl 2008; Chance & Rich 1998; Wang & Liao 2003; Wu & Chen 2007;End-note 21). We can create swap baskets to pick up exposure for the above themes; also swap basketscan be created for the other markets as well, since it could be a good alternative to trading all the otherinvestments.10. Risk and Best Execution MetricsWe can create risk metrics to supplement the information on trade execution in terms of market impact,market risk and optimal timing of transactions (Smithson 1998; Bouchaud & Potters 2003; Wipplinger2007; Christoffersen 2012). It would be helpful to collect information on the market impact of individualtransactions or portfolios based on the timing requirements of implementing the strategies. This marketimpact can be calibrated to the orders, executions and the skill of the executions teams within the firmso that it will capitalize on the trading advantages possessed by the firm. Such information can notonly help with the planning of optimal execution strategies to reduce the market impact and marketrisk of the resulting portfolio, but can be useful to check if the trading strategy is viable; since eventhough it promises to produce good returns but implementing it might eat away the returns. Optimalexecution will be in terms of the number of pieces the overall basket will be broken into and the durationand timing over which to trade the individual pieces. Portfolio risk monitoring will be needed for theindividual strategies and changes in risk levels will need to be monitored for different increments /decrements to the various trading strategies.
To implement these, we would need some derivatives pricing and risk modules, a basic level of which canbe implemented in MATLAB. Since the number of combinations of derivative instruments is huge, only a14rief overview is provided below. Depending on the specifics of the instruments and the market conditions,various strategies can be implemented.1. We can have trades on index options and options on the constituents based on the volatility of theindividual securities and the volatility level of the index and the correlation between individual securitypairs and the average level of correlation in the basket. These are known as volatility and dispersionTrades (Marshall 2009; Meissner 2016; Driessen, Maenhout & Vilkov 2009);2. We predict volatility levels using ARCH / GARCH models and trade instruments that are theoreticallymispriced when compared to the implied volatility levels (Andersen, Bollerslev, Diebold & Labys 2003;Andersen & Bollerslev 1998; Engle & Ng 1993; Xu & Malkiel 2003; Bollerslev, Chou & Kroner 1992;Nelson 1991).3. Put – Call Parity violations (Cremers & Weinbaumv 2010; Klemkosky & Resnick 1979; Finucane 1991),combined with either implied volatilities or based on predicted ARCH / GARCH volatility levels, canbe used to find undervalued or overvalued instruments and appropriate strategies can be constructed.4. Option hedges are possible for basket trades constructed in the delta one section 4.1.5. Structures based on different options, that is structured equity derivatives (Kat 2001) suitable fordifferent Macro, Sector themes and Expectations.
We have looked at a number of delta-one and derivative trading strategies, related market color and pointersto practically implement these investment ideas. The advantages of investing in a moderate amount ofcomputing infrastructure and hiring personnel with technical abilities were illustrated in terms of havinga wider choice of investment alternatives, increased returns and better management of risky outcomes. InKashyap (2018) we discuss in detail a trading strategy, called the bounce basket, for someone to express abullish view on the market by allowing them to take long positions on securities that would benefit the mostfrom a rally in the markets. Given the dynamic nature of the financial markets, it would be practical tohave a feedback loop that changes the parameters of any trading strategy depending on changes in marketconditions (Kashyap 2014a). 15
End-notes
1. Trading Strategy, Wikipedia Link In finance, a trading strategy is a fixed plan that is designed toachieve a profitable return by going long or short in markets. The main reasons that a properlyresearched trading strategy helps are its verifiability, quantifiability, consistency, and objectivity. Forevery trading strategy one needs to define assets to trade, entry/exit points and money managementrules. Bad money management can make a potentially profitable strategy unprofitable.2. Market Color is a word commonly used on trading desks both on the buy side and sell side (End-notes3, 4). It refers to information regarding the financial markets, many times changes in variables that aredeemed relevant to make trading decisions. The sell side provides many such pieces of information tothe buy side, eventually hoping to get trades done on the back of this such material. This could alsobe considered as research done by analysts on both the sell side and the buy side.3. Sell Side, Wikipedia Link Sell side is a term used in the financial services industry. The three mainmarkets for this selling are the stock, bond, and foreign exchange market. It is a general term thatindicates a firm that sells investment services to asset management firms, typically referred to as thebuy side, or corporate entities. One important note, the sell side and the buy side work hand in handand each side could not exist without the other. These services encompass a broad range of activities,including broking/dealing, investment banking, advisory functions, and investment research.4. Buy Side, Wikipedia Link Buy-side is a term used in investment firms to refer to advising institutionsconcerned with buying investment services. Private equity funds, mutual funds, life insurance compa-nies, unit trusts, hedge funds, and pension funds are the most common types of buy side entities. Insales and trading, the split between the buy side and sell side should be viewed from the perspective ofsecurities exchange services. The investing community must use those services to trade securities. The"Buy Side" are the buyers of those services; the "Sell Side", also called "prime brokers", are the sellersof those services.5. Delta One, Wikipedia Link Delta One products are financial derivatives that have no optionality andas such have a delta of (or very close to) one – meaning that for a given instantaneous move in theprice of the underlying asset there is expected to be an identical move in the price of the derivative.Delta one products can sometimes be synthetically assembled by combining options.6. Derivative (finance), Wikipedia Link In finance, a derivative is a contract that derives its value fromthe performance of an underlying entity. This underlying entity can be an asset, index, or interest rate,and is often simply called the "underlying." Derivatives can be used for a number of purposes, including16nsuring against price movements (hedging), increasing exposure to price movements for speculation orgetting access to otherwise hard-to-trade assets or markets.7. Bloomberg, Wikipedia Link Bloomberg L.P. is a privately held financial, software, data, and mediacompany headquartered in Midtown Manhattan, New York City. It was founded by Michael Bloombergin 1981, with the help of Thomas Secunda, Duncan MacMillan, Charles Zegar, and a 30% ownershipinvestment by Merrill Lynch.8. Matlab, Wikipedia Link MATLAB (matrix laboratory) is a multi-paradigm numerical computing en-vironment and proprietary programming language developed by MathWorks. MATLAB allows matrixmanipulations, plotting of functions and data, implementation of algorithms, creation of user interfaces,and interfacing with programs written in other languages, including C, C++, C