The same waves of change described by futurist Alvin Toffler in his books Future Shock, The Third Wave, and Power Shift are surging through Wall Street and impacting the financial industry at all levels. These crashing waves, in some cases, have caused great turmoil and convulsive adjustment. The abruptness of these changes have led some analysts to search for answers in chaos theory and random walk statistics. Rather than suggest that nature has taken over our markets and removed human control, we shall take a fresh look at the problem and then study the forces at work.
For the free enterprise system to work, investors must understand and believe in the effectiveness of the market place. If the market is totally random and beyond interpretation, then it no longer becomes an effective vehicle for channeling resources to where they are needed and can be effectively used. Let's look at the "major waves" that have moved through the financial markets and the impact they have had.
Following the market crash in 1929, the free market system was near
collapse and in great jeopardy. In response to the public outcry and in
an attempt to reestablish confidence in the market system, Congress created
the first great wave of change by passing a series of legislative acts
starting in 1933. The single greatest element in the legislative wave
was the requirement for "disclosure." The six acts that make up the legislative
wave include:
Securities Act of 1933. The "truth in securities" law has two
basic objectives: to provide investors with material, financial, and other
information concerning securities offered for public sale and to prohibit
misrepresentation, deceit, and fraudulent acts and practices in the sale
of securities.
Securities Exchange Act of 1934. This act extended the "disclosure"
doctrine to securities listed and registered for public trading on national
securities exchanges and over-the-counter markets. It also created the
Securities and Exchange Commission (SEC).
Public Utility Holding Company Act of 1935. This statute was
enacted to correct the many abuses that Congressional inquiries had disclosed
in the financing and operation of electric and gas public utility holding
companies.
Trust Indenture Act of 1939. This act was passed to enhance
the reporting and registration requirements for bonds, debentures, notes,
and similar debt securities offered for public sale.
Investment Company Act of 1940. This act required registration
of all companies engaged in the business of investing, reinvesting, and
trading in securities and extended the SEC's regulation over these particular
kinds of companies.
Investment Advisers Act of 1940. This act requires, with certain
exceptions, that persons or firms who engage for compensation in the business
of advising other person or firms about their securities transactions register
with the SEC and conform their activities to the standards designed to
protect investor interest.
Collectively, the disclosure and compliance requirements incorporated in the legislative wave created thousands of jobs for attorneys, accountants, and government bureaucrats. It also resulted in a bonanza for the paper industry. The corporate reporting required under the legislative wave included S1's, S2's, S3's, N1's, N2's, S6's, S18's, annual reports, quarterly reports, 10K's, 18K's, 19K's, 20F's, 10Q's, 8K's, 10C's, 13F's and many other reports and forms. The accumulation of paper in Washington was without precedent. The flood of documents and information almost overwhelmed the regulators, but did little for the investing public. By the time investors received and digested the corporate reports, the information was relatively dated and was probably already reflected in the stock price. This situation led many in the academic community to adopt the Efficient Market Hypothesis, which suggests that all that is to be known about a stock is reflected in its current price. So widely accepted was this theory that the Efficient Market Hypothesis was, and still is, being taught in every graduate business program in the country.
As the paper created by the legislative wave accumulated and the demand for government warehouse space grew at an exponential rate, a new wave emerged. Led by technological breakthrough in computers and telecommunication, American entrepreneurs found a way to implement what the Congressional legislators had hoped to accomplish with their first wave effort: Put accurate, timely information in the hands of the investor. The technology for converting and communicating information around the world at lightning fast speed created a second wave of change that is impacting our current markets in an incredibly dramatic way.
We are now at the point where we are overwhelming investors with data. With or without a computer, investors can, for a few hundred dollars, receive pages of financial data on literally thousands of companies. The business of interpreting all this data is a growth industry. The number of financial advisory newsletters, certified financial analysts, and financial TV commentators is on the rise. Now that you have all this information, what do you do with it? The demand for more interpretation and less raw data is setting the stage for the third wave.
The initial response to the flood of data was to hire more analysts. Most major brokerage firms, Wall Street research firms, and large mutual funds have stables full of financial analysts gathered just to interpret the data now available.
Just as the demand for better data reduction and communications led to the rise of the second wave, so also has the demand for better analysis and interpretation led to the rise of the third wave. We are now experiencing the rise of smart computer systems, and the impact of these systems is going to significantly alter the financial community and the way we manage money.
Now the technology of the smart system is coming to Wall Street. Although the financial community has had experience with computerized buy/sell programs for some time, they were relatively simple computer programs based on mathematical algorithms that weighted baskets of stocks against options, futures, and other derivative product combinations. The new systems will incorporate human knowledge and experience and work on rules of thumb, statistical information theory, and signal detection theory. Eventually, the hybrid neural network systems using fuzzy logic will adjust themselves and actually "learn" as they experience new market conditions. The first of such systems now exists, and the results have been significant.
Discussion of computer smart systems generally centers around the topic of artificial intelligence, known in the industry as AI. The practitioners in the field are known as knowledge engineers and trace their beginning to a conference held at Dartmouth University (Hanover, N.H.) in 1956. The term artificial intelligence is unfortunate, for it is neither artificial nor intelligent. Perhaps a better term to describe the field would be applied intelligence. Regardless of which term is used, AI encompasses several technologies: expert systems, neural networks, fuzzy logic, and genetic algorithms.
EXPERT SYSTEMS
Expert systems are computer programs designed to solve problems that, when solved by humans, require extensive experience and intelligence. They are referred to as rule-based systems or knowledge-based systems because they incorporate the rules of thumb used by experts in a field rather than employing more algorithmic or statistical methods. Expert systems are more capable of dealing with qualitative problems than conventional computer programs. Conventional programs operate on a set of data in a procedural fashion, following the same basic path of logic for each data set and arriving at a correct solution. Expert systems are designed to query users about a problem using a base of knowledge about the general problem domain. Well-designed expert systems ask users only questions relevant to their specific problem. An expert system may fulfill a function that normally requires human expertise, or it may play the role of an assistant to human decision makers. The decision makers may be experts in their own right, in which case the program may justify its existence by improving the decision makers' productivity. Alternatively, human collaborators may be those capable of attaining expert levels of performance given some technical assistance from the program. Successful expert systems are found in many different domains, such as medicine, agriculture, organic chemistry, oil well exploration, equipment malfunction diagnosis, and now portfolio management.
Thus expert systems encode the knowledge of everyday practitioners in some field and use this knowledge to solve problems, instead of using conventional nonadaptive methods derived from computer science or mathematics. RMC Capital Management, a San Diego, California - based money manager, has been using with great success, an expert system since 1987 to guide its investment process. With an annualized rate of return of over 25%, RMC Capital Management has demonstrated the validity of this particular approach.
NEURAL NETWORKS
Some specialized computer programs are inspired by the study of the human nervous system. In a neural network, the unit analogous to the biological neuron is referred to as a processing element. A processing element has many input paths and combines usually by a simple summation, the values on these input paths. The combined input is then modified by a transfer function and passed on to the next-level neuron. A neural network consists of many processing elements joined together in processing layers. The two main phases in the operation of a neural network are learning and recall.
In the first phase, the network adapts or "learns" through a process of adjusting the weights associated with its input pathways. After a network has learned to recognize and categorize its inputs, it can be used in a recall mode to categorize new data even if it is "noisy" and incomplete. Neural computing is suited for pattern matching, pattern recognition, signal processing, and control function synthesis. Because of its ability to detect subtle, hard-to-recognize features, it has significant potential for screening investment alternatives.
A number of trading software systems currently use neural networks. Merrill Lynch (New York, N.Y.) uses a neural net to rate bonds and Fidelity (Boston, Mass.) is managing some of its mutual funds with neural net input. NeuWorld Financial (San Diego, Calif.) has demonstrated the power of nonlinear systems for handling the complex interactions among financial variables.
FUZZY LOGIC
Traditional computer systems and classical set theory have been built on two-state Boolean logic. Things are either true or false, plus or minus, present or absent, or ones and zeros. Unfortunately, things in the real world are never that clear cut. Humans use language to reason things out, and our language is far from precise. What does it mean when we say someone is tall or a company has strong earnings? To handle this type of real-world problem, a new form of logic/mathematics has been derived called fuzzy set theory. It forms the basis for fuzzy logic and possibility theory and improves the reasoning ability of expert systems and neural networks. The incorporation of this type of logic is impacting engineering system designs, and we now have cameras focused, trains controlled, and TV's tuned by fuzzy systems. Certainly, few problems are fuzzier than market economics. We can envision numerous applications in the financial industry.
A number of financial variables involved in financial decisions are, by their very nature, fuzzy. We talk about high price/earnings stocks. Well, what is high? The definition of high will change depending on the growth rate of the company, the industry sector the company is in, whether the market is moving higher or lower, whether interest rates are high or low, and many other factors. To handle the fuzzy nature of this and other financial data, we must resort to fuzzy set theory.
GENETIC ALGORITHMS
This term describes a Darwinian approach to the development of a system. When an evolutionary process takes place, the resulting organism or system is different from its parents and is possibly more or less "fit." If it is more fit, it will survive and pass on its characteristics. If less fit, it is less likely to survive to pass on its characteristics. Over many generations, this process will increase the average "fitness" of the species or system.
Markets continue to change, and our decision support systems must also change. A recent Association for Investment Management and Research (Charlottesville, VA) conference in Boston, Mass., was highlighted by a number of presentations that provided statistical data on the changing markets. Systems built prior to the market crash of 1987 are not performing well new. Decision support systems must be able to change or evolve.
OTHER TECHNOLOGIES
As more people are educated in multiple disciplines, further cross fertilization of ideas and many scientific theories, applications, and technologies currently used in other field will find their way into the financial field. Some of the more obvious ones include statistical information theory, signal detection theory, linear programming, decision theory, and advanced process modeling techniques such MicroSaint by Micro Analysis and Design (Boulder, Colo.). Most, it not all, of these technologies will have an impact on how we manage money in the future.
This third-wave technology has introduced a new dimension to the financial field, a dimension I shall call "financial cybernetics." Future graduate programs in business and finance will have to include courses in business or financial cybernetics if they are to stay competitive in this rapidly changing world.
Thomas E. Berghage is president and CEO of NeuWorld Financial, a San Diego, California based research and development corporation specializing in advanced AI systems for the financial industry.