The past 30 years have seen a profound transformation in the world's financial markets fueled by the globalization of markets and the dramatic advances in computer and telecommunication technology. As a result, there has been a proliferation in the types and complexity of securities; witness the explosive growth in the derivative and asset-backed securities markets. Among the other features of the changing nature of the financial landscape is the flood of new money into the securities markets as a result of the growth in pension funds, 401(K)s, IRA's etc. and the movement of control of investment funds from the individual to the institutional money manager.
Paralleling the remarkable transformation of the financial markets has been the development of what is called by some the "new science of finance," an evolving interdisciplinary approach combining financial research, statistics and other mathematical disciplines, physics, computer science, biology, behavioral and cognitive science.
Some of the early quantitative methods included the successful identification by Yale mathematics professor Benoit Mandelbrot of patterns in cotton prices in the 1960s. From this and subsequent investigations, Mandelbrot developed the theory of fractals, which applies to such divergent phenomena as turbulence in liquids, the structure of plants, and the behavior of markets.
By the 1980s, there was mounting evidence that predictive systems could be built and successfully deployed:
* Work by Mandelbrot and others demonstrated fractal patterns in financial markets. Patterns were found and exploited governing the volatility of foreign exchange markets. It was found, for example, that volatility was dependent in part on who was participating in the market at a given time. For instance, volatility dropped during the time that Asian traders went to lunch and increased when New York and London were both trading.
* Robert Engle found that volatility was not random but that periods of high and low volatility last longer than the traditional random models had predicted. Engle discovered "clustered" models of volatility that had predictive value.
* In a 1990 study, Andrew Lo and C.A. Mackinlay found that a positive return in weekly stock indices were positively correlated with positive returns during the following week.
* In a 1991 study by the Santa Fe Institute, which put popular technical analysis techniques to a rigorous test using 90 years of data from the Dow Jones Index, it was found that both the moving average rule and trading-range break rule worked quite well.
These and other developments increased the interest in quantitative methods. Today, predictive financial models are being developed using a variety of techniques, including fuzzy logic, neural nets, genetic algorithms, Markov models, fractal methods, and clustering techniques. Neural nets have been used by Citibank, Credit Suisse, Advanced Investment Technology (owned by State Street Global Advisors), Nikko Securities, Normura Securities, Morgan Stanley, Bear Stearns, Shearson Lehman Hutton, and Fidelity Management & Research.
First Quadrant, which manages over 27 billion dollars in assets (including $2.2 billion in long-term strategies), is using genetic algorithms for research purposes in a variety of their financial services. The Renaissance Technologies hedge fund, run by mathematician James Simons, although a relatively small fund (about $2 billion), has reported a 20% annual return for the past ten years using all quant methods. Another fund run by the "Prediction Company," which uses nonlinear time series methods developed by former physicists from Los Alamos, has also had very positive results.
The biggest contemporary impact of quant methods is as a tool used in combination with human decision making. The International Association of Financial Engineers, the leading professional association for the growing field of financial engineering, conducted a study of the usage of quantitative techniques in financial firms. The survey was sent to 200 of the largest investment banks, broker/dealers, insurance companies, and other financial institutions around the world. The focus of the study was the importance of the role of quantitative techniques and the use of mathematics talent in implementing financial management techniques. Of the 50 respondents, 40% reported a heavy dependence on quantitative methods and 75% cited a moderate or higher dependence. 76% of the respondents noted that the proportion of quantitative professionals in their firm has increased over the past 15 years with 71% expecting that the proportion will continue to increase. Another indication of the increasing excitement and potential surrounding quantitative methods is that a large proportion of the doctorate graduates in theoretical physics from universities like Harvard and Stanford are going into financial research.
There are a number of trends indicating that quantitative methods will be of enormous importance in the years ahead:
* The rapidly increasing availability of online financial information, which is necessary to train self-organizing methods such as genetic algorithms and neural nets, is growing very rapidly.
* Patterns in financial trading data representing arbitrage opportunities are often too subtle for human analysts to detect, but very powerful statistical techniques (combined with application of significant computer processing) can detect these patterns and develop strategies for exploiting them.
* Computational power is growing exponentially, which is another important ingredient for complexity theory methods. There is also growing availability of dedicated parallel devices for neural net connection calculations and genetic algorithms.
* Advances in quantitative techniques are leveraged by the enormous size of the market, now estimated at over $20 trillion. According to Thomson Financial, of the $20.9 trillion of total market capitalization in 1997, $12 trillion was controlled by institutional money managers, and, in a recent publication, Instinet Corporation (a subsidiary of Reuters Group PLC) states that the value of all holdings in the U.S. "increased from $849.9 billion in 1969 to $14.6 trillion in 1998."
* The four largest U.S. institutional investors: Fidelity ($695 billion under management), Barclays ($615 billion under management), Merrill Lynch ($501 billion under management) and State Street ($490 billion under management) account for $2.3 trillion or over 15% of total securities in the U.S. These firms are in highly competitive environments; their clients are able to track their performance and move their funds readily to the best performers. Being able to gain even a slight advantage by obtaining superior technology when leveraged by the large sums under management would have very significant value. Fidelity, in its own description of itself, states that, "this year alone, Fidelity will invest $500 million dollars in hardware, software, and systems that enable us not only to analyze and research virtually all the world's markets, but to provide customers with the most up-to-the-minute information they need to help them make sound financial decisions." As huge amounts of investment are at stake, we believe that these firms are prepared to move quickly to maintain and increase market share.
Examples of quant funds include the following: In 1997, D.E. Shaw & Co., using quantitative methods, accounted for up to 5% of the trading on the NYSE; as of mid-1998, the quant firm BNP/Cooper Neff Advisors accounts for 4 to 6 percent of the total on the NYSE and 6 to 10 percent of the volume of a number of European exchanges volume; First Quant's managed assets grew from $11 billion in 1997 to $27 billion in the final quarter of 1998; and, according to Andrew W. Lo "as much as 20% of hedge funds, which control some $295 billion, are quantitatively oriented."
Quant funds are making significant inroads into the institutional investor community. In a 1998 Wall Street Research report, Ashton Technology Group, Inc. states that quant funds (along with index funds) represent the fastest growing institutional investor groups. Industry observers estimate that quant funds control approximately 5% of the $20 trillion in the market, and are growing rapidly. Goldman Sachs was reported to have purchased a quant technology company for $600 million.
Source: http://www.fatkat.com/market.html
Paralleling the remarkable transformation of the financial markets has been the development of what is called by some the "new science of finance," an evolving interdisciplinary approach combining financial research, statistics and other mathematical disciplines, physics, computer science, biology, behavioral and cognitive science.
Some of the early quantitative methods included the successful identification by Yale mathematics professor Benoit Mandelbrot of patterns in cotton prices in the 1960s. From this and subsequent investigations, Mandelbrot developed the theory of fractals, which applies to such divergent phenomena as turbulence in liquids, the structure of plants, and the behavior of markets.
By the 1980s, there was mounting evidence that predictive systems could be built and successfully deployed:
* Work by Mandelbrot and others demonstrated fractal patterns in financial markets. Patterns were found and exploited governing the volatility of foreign exchange markets. It was found, for example, that volatility was dependent in part on who was participating in the market at a given time. For instance, volatility dropped during the time that Asian traders went to lunch and increased when New York and London were both trading.
* Robert Engle found that volatility was not random but that periods of high and low volatility last longer than the traditional random models had predicted. Engle discovered "clustered" models of volatility that had predictive value.
* In a 1990 study, Andrew Lo and C.A. Mackinlay found that a positive return in weekly stock indices were positively correlated with positive returns during the following week.
* In a 1991 study by the Santa Fe Institute, which put popular technical analysis techniques to a rigorous test using 90 years of data from the Dow Jones Index, it was found that both the moving average rule and trading-range break rule worked quite well.
These and other developments increased the interest in quantitative methods. Today, predictive financial models are being developed using a variety of techniques, including fuzzy logic, neural nets, genetic algorithms, Markov models, fractal methods, and clustering techniques. Neural nets have been used by Citibank, Credit Suisse, Advanced Investment Technology (owned by State Street Global Advisors), Nikko Securities, Normura Securities, Morgan Stanley, Bear Stearns, Shearson Lehman Hutton, and Fidelity Management & Research.
First Quadrant, which manages over 27 billion dollars in assets (including $2.2 billion in long-term strategies), is using genetic algorithms for research purposes in a variety of their financial services. The Renaissance Technologies hedge fund, run by mathematician James Simons, although a relatively small fund (about $2 billion), has reported a 20% annual return for the past ten years using all quant methods. Another fund run by the "Prediction Company," which uses nonlinear time series methods developed by former physicists from Los Alamos, has also had very positive results.
The biggest contemporary impact of quant methods is as a tool used in combination with human decision making. The International Association of Financial Engineers, the leading professional association for the growing field of financial engineering, conducted a study of the usage of quantitative techniques in financial firms. The survey was sent to 200 of the largest investment banks, broker/dealers, insurance companies, and other financial institutions around the world. The focus of the study was the importance of the role of quantitative techniques and the use of mathematics talent in implementing financial management techniques. Of the 50 respondents, 40% reported a heavy dependence on quantitative methods and 75% cited a moderate or higher dependence. 76% of the respondents noted that the proportion of quantitative professionals in their firm has increased over the past 15 years with 71% expecting that the proportion will continue to increase. Another indication of the increasing excitement and potential surrounding quantitative methods is that a large proportion of the doctorate graduates in theoretical physics from universities like Harvard and Stanford are going into financial research.
There are a number of trends indicating that quantitative methods will be of enormous importance in the years ahead:
* The rapidly increasing availability of online financial information, which is necessary to train self-organizing methods such as genetic algorithms and neural nets, is growing very rapidly.
* Patterns in financial trading data representing arbitrage opportunities are often too subtle for human analysts to detect, but very powerful statistical techniques (combined with application of significant computer processing) can detect these patterns and develop strategies for exploiting them.
* Computational power is growing exponentially, which is another important ingredient for complexity theory methods. There is also growing availability of dedicated parallel devices for neural net connection calculations and genetic algorithms.
* Advances in quantitative techniques are leveraged by the enormous size of the market, now estimated at over $20 trillion. According to Thomson Financial, of the $20.9 trillion of total market capitalization in 1997, $12 trillion was controlled by institutional money managers, and, in a recent publication, Instinet Corporation (a subsidiary of Reuters Group PLC) states that the value of all holdings in the U.S. "increased from $849.9 billion in 1969 to $14.6 trillion in 1998."
* The four largest U.S. institutional investors: Fidelity ($695 billion under management), Barclays ($615 billion under management), Merrill Lynch ($501 billion under management) and State Street ($490 billion under management) account for $2.3 trillion or over 15% of total securities in the U.S. These firms are in highly competitive environments; their clients are able to track their performance and move their funds readily to the best performers. Being able to gain even a slight advantage by obtaining superior technology when leveraged by the large sums under management would have very significant value. Fidelity, in its own description of itself, states that, "this year alone, Fidelity will invest $500 million dollars in hardware, software, and systems that enable us not only to analyze and research virtually all the world's markets, but to provide customers with the most up-to-the-minute information they need to help them make sound financial decisions." As huge amounts of investment are at stake, we believe that these firms are prepared to move quickly to maintain and increase market share.
Examples of quant funds include the following: In 1997, D.E. Shaw & Co., using quantitative methods, accounted for up to 5% of the trading on the NYSE; as of mid-1998, the quant firm BNP/Cooper Neff Advisors accounts for 4 to 6 percent of the total on the NYSE and 6 to 10 percent of the volume of a number of European exchanges volume; First Quant's managed assets grew from $11 billion in 1997 to $27 billion in the final quarter of 1998; and, according to Andrew W. Lo "as much as 20% of hedge funds, which control some $295 billion, are quantitatively oriented."
Quant funds are making significant inroads into the institutional investor community. In a 1998 Wall Street Research report, Ashton Technology Group, Inc. states that quant funds (along with index funds) represent the fastest growing institutional investor groups. Industry observers estimate that quant funds control approximately 5% of the $20 trillion in the market, and are growing rapidly. Goldman Sachs was reported to have purchased a quant technology company for $600 million.
Source: http://www.fatkat.com/market.html
