T h e Q u a n T C l a s s r o o m b y aT T i l i o m e u C C i

Differences and Commonalities between the Two Areas of Quantitative Finance

The world of derivatives pricing and the world of risk and portfolio management are different, but not so different.

versus :

here exist two separate branches of finance that The goal of derivatives pricing is to determine the fair price require advanced quantitative techniques: the of a given security in terms of more liquid securities for which area of derivatives pricing, which is tasked with the price is determined by the law of supply and demand. Exextrapolating the present; and the area of amples of securities being priced are exotic options, mortgagequantitative risk and portfolio management, which backed securities, convertible bonds, structured products, etc. is tasked with modeling the future. Once a fair price has been determined, the sell-side trader We briefly trace the history of these two branches of quanti- can make a market on the security. Therefore, derivatives pricGoal: "extrapolate the present" tative finance, highlighting their different goals and challenges. ing is a complex extrapolation exercise to define the current Then we provide an overview of their areas of intersection: market value of a probability Q is then used by the sell-side Environment: risk-neutral security, which the notion of risk premium; the stochastic processes used, often community. Processes: continuous-time martingales under different names and assumptions in the and in the Quantitative derivatives pricing was initiated by Bachelier Dimension: world; the numerical methods utilized to simulate those pro- (1900)low the introduction of the most basic and most influwith cesses; hedging; and statistical arbitrage. of calculus, the Brownian motion, and its applications Tools: ential Ito processes, PDEs to pricing of Challenges:the calibrationoptions. The theory remained dormant until Derivatives Pricing: the World Merton (1969) and Black and Scholes (1973) applied the secBusiness: sell-side In this section, we provide a brief overview of the world of ond most influential process, the geometric Brownian motion, derivatives pricing. Refer to Figure 1 below for a summary. to option pricing. The next important step was the fundamental theorem of Figure 1: pricing by Harrison and Pliska (1981), according to which the The Q world of derivatives pricing Figure 1: The Q World of Derivatives Pricing suitably normalized current price P0 of a security is arbitrage Goal: extrapolate the present free, and thus truly fair, only if there exists price P0 of Pliska (1981), according to which the suitably normalized current a stochastic process Environment: risk-neutral probability arbitrage-free,Pand thus trulyexpected value which describes its future evolut with constant fair, only if there exists a stochastic a security is tion: process Pt with Processes: continuous-time martingales constant expected value which describes its future evolution:
Dimension: Tools: Challenges: Business: low

P0 = E {Pt } ,

t 0.

(1)

(1)

Ito calculus, PDEs A process satisfying (1) is called a "martingale". A martingale does not reward A process satisfying (1) is called a martingale. A martingale

risk. Thus the probability of the reward risk. Thus, theprice process is called calibration does not normalized security probability of the normalized

www.garp.org

"risk-neutral" and is typically denoted by the blackboardrisk-neutral and The security price process is called font letter "Q". is typically sell-side relationship (1) must hold for all times t: therefore the processes used for derivdenoted by the blackboard font letter . The relationship atives pricing are naturally set in continuous time. The quants who operate in the Q world of derivatives pricing are specialists with deep knowledge of the specic products they model. Securities are priced individually, and thus the problems in B R U Aworld 2 0 1low-dimensional F E the Q R Y are 1 RISK PROFESSIONAL in nature. Calibration is one of the main challenges of the Q world: once a continuous-

41

T h e Q u a n T C l a s s r o o m b y aT T i l i o m e u C C i

T h e Q u a n T C l a s s r o o m b y aT T i l i o m e u C C i

(1) must hold for all times t: therefore, the processes used for derivatives pricing are naturally set in continuous time. The quants who operate in the world of derivatives pricing are specialists with deep knowledge of the specific products they model. Securities are priced individually, and thus the problems in the world are low-dimensional in nature. Calibration is one of the main challenges of the world: once a continuous-time parametric process has been calibrated to a set of traded securities through a relationship such as (1), a similar relationship is used to define the price of new derivatives. The main quantitative tools necessary to handle continuoustime -processes are Itos stochastic calculus and partial differential equations (PDEs). Throughout the past decades, these advanced techniques have attracted mathematicians, physicists and engineers to the field of derivatives pricing. Risk and Portfolio Management: the World In this section, we provide a brief overview of the world of risk and portfolio management. Refer to Figure 2 below for a summary.

Figure 2: The P World of Risk and Portfolio Management

Goal: Environment: Processes: Dimension: Tools: Challenges: Business: model the future real probability discrete-time series large multivariate statistics estimation buy-side

ment started with the mean-variance framework of Markowitz (1952). Next, breakthrough advances were made with the Capital Asset Pricing Model (CAPM) and the Arbitrage Pricing Theory (APT) developed by Treynor (1962), Mossin (1966), Sharpe (1964), Lintner (1965) and Ross (1976). The above theories provide tremendous insight into the markets, but they assume that the probability distribution is known. In reality, the probability distribution must be estimated from available information. A major component of this information set is the past dynamics of prices and other financial variables, which are monitored at discrete time intervals and stored in the form of time series. Estimation represents the main quantitative challenge in the world of risk and portfolio management. The analysis of the time series requires advanced multivariate statistics and econometric techniques. Notice that in risk and portfolio management it is important to estimate the joint distribution of all the securities in that market, and thus securities cannot be considered individually as in the world of derivatives pricing. Therefore, dimension reduction techniques such as linear factor models play a central role in the world. To address the above issues, in recent years a new breed of quants, the -quants, has started to populate the financial industry, and more -quants are being trained in the same masters programs that were originally designed to prepare -quants. Commonalities Between and From a comparison of Figure 1 and Figure 2, it appears that the and the world of quantitative finance are very different. In reality, commonalities between these two worlds abound and interactions occur frequently in different areas, as we proceed to illustrate. Risk Premium Mathematically, the risk-neutral probability and the real probability associate different weights to the same possible outcomes for the same financial variables. The transition from one set of probability weights to the other defines the so-called risk-premium. Knowledge of the risk-premium allows us in principle to switch from one world to the other. Unfortunately, the correct estimation of the risk premium is a challenging task. Stochastic Processes Stochastic processes are the building blocks of any quantitative model, both in the world and in the world. Although -

quants focus on continuous risk-neutral processes and -quants focus on discrete-times processes, the same models are used in both areas, possibly under different assumptions and names. Below, we provide a brief overview of such processes and of their main features, which we summarize in Figure 3. The interested reader can find in Meucci (2009a) a more exhaustive overview, a thorough theoretical discussion, an empirical analysis, fully documented code, and further references.

model, is the most popular stochastic volatility model to price derivatives. Numerical Methods The theoretical stochastic processes previously discussed must be implemented in practice. In order to do so, the most popular numerical techniques are trees and Monte Carlo simulations. Trees represent a process as a sequence of an ever-expanding set of potential outcomes: the state of the world today will give rise to multiple possible outcomes tomorrow; each of these in turn will give rise to multiple possible outcomes the day after tomorrow, and so on. As we see, with trees the number of potential outcomes grows with the horizon. With Monte Carlo simulations, the number of possible outcomes that represent a stochastic process, known as paths, is kept constant throughout the evolution of the process. The computationally more costly trees are used when it is important to make decisions along the trajectory of the stochastic process, whereas Monte Carlo is used when only the process distribution is required. Therefore, in the world of risk and portfolio management, trees are used to design dynamic strategies, whereas Monte Carlo is used for risk monitoring purposes such as value-at-risk computations. In the world, trees are used for instance to price American options, which can be exercised earlier than at expiry, whereas Monte Carlo is used to price Asian options i.e., options on the average price of an underlying instrument over a pre-specified period of time. Hedging Hedging is a clear example where the world and the world interact directly. Hedging aims at protecting the future p&l of a given position from a set of risk factors. Therefore, hedging is a -world concept. In order to determine the amounts of the hedging instruments to buy or sell, we need to compute the sensitivity of the given position and of the hedging instruments to those risk factors. Such sensitivities are known as the Greeks. The most famous Greek is the delta of an option written on a given security, which is the sensitivity of the option to the underlying security. The delta of an option tells the trader how much underlying to sell in order to protect the option from swings of the underlying. The Greeks are computed using pricing models from the

Figure 3: Fundamental Stochastic Processes for the and the Worlds

Base case: Autocorrelation: Vol. clustering: Discrete-Time random walk ARMA GARCH Continuous-Time Levy (Brownian, Poisson) Ornstein-Uhlenbeck

stochastic volatility subordination

Risk and portfolio management aims at modeling the probability distribution of the market prices at a given future investment horizon. This real probability distribution of the market prices is typically denoted by the blackboard font letter , as opposed to the risk-neutral probability used in derivatives pricing. Based on the distribution, the buy-side community makes decisions on which securities to purchase in order to improve the prospective profit-and-loss profile of their positions considered as a portfolio. The quantitative theory of risk and portfolio manage-

The most fundamental discrete-time process is the random walk, which is the cumulative sum of invariants i.e., variables that behave independently and identically across time. The random walk is the baseline assumption to model interest rates or the log-price of stocks in risk and portfolio management. In continuous time, random walks become Levy processes. The Brownian motion, which is the most notable instance of a Levy process, is the baseline process for option pricing. Similarly, the Poisson process, which is the second simplest Levy process, pervades the pricing of credit products. A second class of processes models autocorrelation. Autocorrelation occurs when financial series are not the sum of independent increments. The standard tool to model autocorrelation in discrete time are auto-regressive-moving-average (ARMA) processes, a favorite of buy-side econometricians. The continuous-time version of ARMA processes are Ornstein-Uhlenbeck and related processes. In particular, two forms of the Ornstein-Uhlenbeck process namely, the model by Vasicek (1977) and the model by Cox, Ingersoll and Ross (1985) (CIR) represent the base case sell-side processes to price bonds. A third class of processes models volatility clustering: periods of increased activity tend to occur in bulks. In discrete time, GARCH and generalizations thereof capture this feature for the buy-side audience. On the sell-side, volatility clustering is modeled in two ways: stochastic volatility and subordination. In particular, the model by Heston (1993), derived from the CIR

42

RISK PROFESSIONAL F E B R U A R Y 2 0 1 1

www.garp.org

www.garp.org

F E B R U A R Y 2 0 1 1 RISK PROFESSIONAL

43

T h e Q u a n T C l a s s r o o m b y aT T i l i o m e u C C i

The world of derivatives pricing has also moved into the world of risk and portfolio management in the area of statistical arbitrage algorithms.
world and then applied in the -world for hedging. Interestingly, those very same -world pricing models can be derived based on the -world concept of hedging. Statistical Arbitrage The world has also moved into the world in the area of statistical arbitrage algorithms. The general steps of this interaction are as follows: First, models are used to identify misalignments among security prices today. Second, one assumes that the misaligned security prices will eventually converge to the prices predicated by the models. Consequently, a potential expected return in the real world, or alpha, is identified as the difference between the -predicated prices and the current misaligned prices. Third, if the alpha is positive, a long position is set up i.e., the misaligned securities are bought; if the alpha is negative, a short position is set up i.e., the misaligned securities are sold. We sketch this in Figure 4 (below); refer to Meucci (2009b) for more details.

REFERENCES Bachelier, L., 1900. Theorie de la speculation, Annales Scientifiques de lEcole Normale Superieure 3, 2186. Black, F., and M. S. Scholes, 1973. The pricing of options and corporate liabilities, Journal of Political Economy 81, 637654. Cox, J. C., J. E. Ingersoll and S. A. Ross, 1985. A theory of the term structure of interest rates, Econometrica 53, 385407. Harrison, J. M. and S. R. Pliska, 1981. Martingales and stochastic integrals in the theory of continuous trading, Stochastic Processes and their Applications 11, 215260. Heston, S. L., 1993. Closed-form solution for options with stochastic volatility with applications to bond and currency options, The Review of Financial Studies 6, 327343. Lintner, J., 1965. The valuation of risky assets and the selection of risky investments in stock portfolios and capital budgets, Review of Economics and Statistics 47, 1337. Markowitz, H. M., 1952. Portfolio selection, Journal of Finance 7, 7791. Merton, R. C., 1969. Lifetime portfolio selection under uncertainty: The continuous case, Review of Economical Statistics 51, 247257. Meucci, A., 2009a. Review of discrete and continuous processes in finance: Theory and applications. Working paper, available at http://ssrn.com/abstract=1373102. Ibid, 2009b. Review of statistical arbitrage, cointegration, and multivariate Ornstein-Uhlenbeck. Working paper, available at http:// ssrn.com/abstract=1404905. Mossin, J., 1966. Equilibrium in a capital asset market, Econometrica 34, 768783. Ross, S., 1976. The arbitrage theory of capital asset pricing, Journal of Economic Theory 13, 341360. Sharpe, W. F., 1964. Capital asset prices: A theory of market equilibrium under conditions of risk, Journal of Finance 19, 425442. Treynor, J. L., 1962. Toward a theory of market value of risky assets. Unpublished manuscript. Vasicek, O., 1977. An equilibrium characterization of the term structure, Journal of Financial Economics 5, 177188.

Figure 4: Statistical Arbitrage Short Position

Attilio Meucci is the chief risk officer at Kepos Capital LP. He runs the 6-day Advanced Risk and Portfolio Management Bootcamp, see www.symmys.com. The author is grateful to Peter Carr, Emanuel Derman, Gianluca Fusai, Beibi Garli, Adam Lichtenstein and an anonymous referee.

44

RISK PROFESSIONAL F E B R U A R Y 2 0 1 1

www.garp.org