Journal of Technical Analysis (JOTA). Issue 09 (1980, November)

Market Technicians Association

JOURNAL Issue 9 November 1980

- MARKET TECHNICIANS ASSOCIATION JOURNAL

Issue 9

November 1980 -

Published by : Market Technicians Association 70 Pine Street

New York, New York 10005

Copyright 1980 by Market Technicians Association

-l-

intentionally blank

-2-

Market Technicians Association Journal

Editor: Anthony W. Tabell Delafield , Harvey, Tabell 909 State Road Princeton, New Jersey 08540

Associate Editor: S hary Anaya Delafield , Harvey, Tabell

Thanks to the following MTA members and subscribers for their part in the creation of this issue:

Bernadette M. Bartels Jack D . Becker William DiIanni Walter L. Eckardt , Jr. Bernie Fremerman Morton Jacobs John R. McGinley, Jr. David L. Upshaw

-3-

THE MTA JOURNAL

EDITORIAL 9

THE OPTION MARKET -- EARLY TRENDS Bernadette M. Bartels

13

Bernadette M. Bartels, MTA President 1977-1978, has pioneered in the collection and technical analysis of options data since the very inception of listed options trading. She has been instrumental in persuading the Options Clearing Corporation and the various exchanges to provide data which the stock market technician can utilize. She has developed a number of techniques for the analysis of options data and has been instrumental in assisting others in the development of such techniques.

Despite the expansive growth in options trading, technical analysis of options data remains in its infancy. In this article, Miss Bartels gives an overview of the options trading mechanisms, the data currently available, and outlines some analytical techniques which have proved helpful.

A RANDOM WALK THROUGH RANDOM NUMBERS Bernie Fremerman and Morton Jacobs, M.D.

35

The random walk hypothesis has been the bete noire of technicians for the past decade. Yet, in one sense, there is no such thing as randomness. Random number series are reqularly generated by computer, and the programmer who writes such a random number generator will be able to predict the next number in the series with 100 percent accuracy. This phenomenon and its implications are examined in some detail by Fremerman and Jacobs who choose what should be a “random” series, numbers in a Kansas City telephone book, and find some surprising evidence of non-randomness.

BUY SIDE, SELL SIDE, ALL AROUND THE TOWN David L. Upshaw

The MTA membership roles can, by and large, be divided into two categories : Technicians who work on the “buy side” of the street --- for money management firms who distribute commissions --- and those who work on the “sell side” --- for the brokers who receive those commissions. While we all consider ourselves professionals, sharing the same discipline, our point of view is natu- rally colored by the side of the Street on which we earn our living.

43

-4-

Issue No, 9 November, 1980

David Upshaw , MTA’s President in 1976-1977, possesses a unique perspective. He began his career on the “buy side ,‘I switched for an eight-year period, and recently has returned. His article draws on his unique experience and offers some interesting observations about the dichotomy in our profession.

CORPORATE NEWS REPORTS AND SECURITY RETURNS Walter L. Eckardt, Jr. and Jack D. Becker

Technicians have generally agreed on the concept that fundamental news tends to be reflected in the market place, often before the fact. The authors conduct a rigorous examination of the behavior of individual stocks in response to known news events using both previously developed, rudimentary techniques of analysis and more sophisticated ones growing out of the CAPM. All methodologies seem to produce similar results showing, among other things, the market’s tendency to react to good and bad news prior to its being reported. There remains, however, some difficulty in proving a direct relationship between specific information items and stock performance.

THE SHORT-TERM TRADING (ARMS) INDEX REVISITED John R. McGinley, Jr.

The Short-term Trading Index (the advance/decline ratio divided by the up volume/down volume ratio) is probably the most widely analyzed of any of the modern stock market indicators. Its inventor, Dick Arms, has twice spoken before the MTA, and various members have done extensive work on the subject, including your editor who discoursed for an hour at the 1979 seminar on an extensive computer study thereof. John McGinley, the author of this article, has also written previous pieces on the subject for this publication and for the Newsletter.

Nonetheless, the article demonstrates that, if one is willing to spend the time, it is always possible to come up with something new. John explains his simplified calculation of a logarithmic moving average and discusses results achieved by plotting the data on a point-and- figure basis.

ANOTHER POINT ABOUT POINT-AND-FIGURE David L. Upshaw

Point-and figure charts have long had their adherents. However, few technicians have been terribly creative in creating variations on the standard theme of one-point unit charts on individual stocks. One

47

65

69

-5-

INDEX (Continued)

-

such possible variation, long-term charts of group averages, based on monthly average prices, is discussed here together with some of its unique advantages and limitations.

BOOK REVIEW William Dilanni

75

CYCLES Dick A. Stoken McGraw-Hill Book Company

-6-

MARKET TECHNICIANS ASSOCIATION

MEMBERSHIP and SUBSCRIPTION INFORMATION

REGULAR MEMBERSHIP - $50 per year plus $10 one-time application fee.

Receives the Journal, the monthly MTA Newsletter, invitations to all meetings, voting member status and a discount on the Annual Seminar Fee. Eligibility requires that the emphasis of the applicant’s professional work involve technical analysis.

SUBSCRIBER STATUS - $50 per year plus $10 one-time application fee.

Receives the Journal and the MTA Newsletter, which contains shorter articles on technical analysis, and the subscriber received special announcements of the MTA meetings open to The New York Society of Security Analysts and/or the public, plus a discount on the Annual Seminar Fee.

ANNUAL SUBSCRIPTION TO THE MTA.JOURNAL - $35 per year.

SINGLE ISSUES OF THE MTA JOURNAL (including back issues)

are available for $15 to regular members or subscribers $15 to non-members and non-subscribers

The Market Technicians Association Journal is scheduled to be published three times each. fiscal year, in approximately November, February, and May.

An Annual Seminar is heldeach spring.

Inquiries for Membership should be directed to:

Fred R. Gruber, V.P. United Jersey Bank 210 Main Street Hackensack, New Jersey 07602

-7-

intentionally blank

-8-

Editorial

ON WITCH DOCTORS AND QUANTS

Taking over the editorship of this journal from Bill DiIanni is, in the familiar phrase, a tough act to follow. Bill’s efforts, we think, raised the Journal to a standard of excellence few of us had believed possible in an organization as small as our own. It is our profound hope that these standards will not deteriorate too severely under the present editorship.

One of Bill’s policies, and one which we intend, without apology, to continue, was the acceptance and publication of articles with a heavy statistical flavor. While large doses of Greek letters may be a bit off- putting to many of us, it is our strong feeling that the formal documenta- tion results in terms of statistical theory constitute an entirely proper field of activity for the market technician.

It is necessary to face the fact that we technicians are now not the only people looking at stock prices. We have been joined by a group of colleagues who have been dubbed by Institutional Investor magazine as “quants , ” this appelation being short for “quantitative,” and generally taken to mean those analysts who apply reasonably complex mathematical techniques to stock price analysis. Our present relationship with these individuals seems similar in some ways to the relationship between our friends and allies, the British and the French. These two nations spent a few hundred y.ears of early European history fighting wars with each other, and still find themselves separated by substantial idiosyncratic differences such as the relative merits of snails and brussel sprouts, and certain matters relating to personal hygiene. Nonetheless, in recent years, the inhabitants of the two sides of the Channel have been able to ignore their deep-seated historical differences and their current variations in character traits, joining together in such diverse enterprises as two world wars and the Common Market. Such a rapprochement, we think, is the future of the relationship between ourselves and quantitative theorists.

-9-

The historical animosities are, without a doubt, real. The earliest .quantitative work was concentrated in the riper pastures of the Groves of Academe and centered almost exclusively on the extreme interpretations of the Random Walk hypothesis. This emphasis, since it offended large numbers of people, was sure to generate publicity, and publicity is the surest means for converting upwardly- mobile instructors into full professors. This particular historical phase is, we think, now behind us.

The Random Walk hypothesis has now surfaced in a number of new manifestations such as the efficient market hypothesis and the capital asset pricing model. It has also, whether we like it or not, left the ivory towers and arrived at a desk down the hall. We do not think that living with it will be an impossible task.

It seems to us that, while there is plenty of room for disagreement between our two camps on such matters as, for example, just how “efficient” the stock market really is, there does not exist the same huge gulf between techncial and quantitative theory that existed between the technical fraternity of a couple of decades ago and the early academics, a gulf typified by Paul Cootner’s description of technicians as witch doctors. We think that, given a few ground rules we will all be able to share the same turf, and a few suggestions for such ground rules are herewith offered.

For quantitative analysts, our first suggestion is that they learn to speak English, a process which, we think, may already be under way. We have personally encountered nothing in the mathematical disciplines that could not be explained - - - at least conceptually - - - without reference to algebraic symbols, and even without using the words “stochasticl’ and “heristic.ll On the other hand, mathematical demonstration has its place in the exposition of any technique, and the language of mathematics, even unto the Greek letters, has evolved over the years simply because that language is useful in such exposition. Despite proverbs about old dogs and new tricks, it might even be useful it the technician learned some of this language. If our children manage to swallow binary numbers in the fifth grade, we do not need to choke on a few alphas and betas.

-lO-

We, for our own part, should avoid know-nothingism i It is all very well to carp at pointy-headed academics and suggest that only old-time, seat-of-the-pants chartists have the common sense to detect any meaning in the stock market. This is an apoealing posture for those who are too lazy to learn, but it probably just ain? so.

We are all here, it seems to us, because a number of basic facts remain true about the real world. We live, in a capitalistic society, from which it necessarily follows that private decisions must be make about the investment of funds in capital assets. Such decisions need to be made with all of the accumulated knowledge and with the best techniques at our disposal. It is now generally agreed that one such technique is the analysis of past stock price data, which is precisely the point where all of us, witch doctors and quants, fit into the picture. We would do well to get on with the task.

ANTHONY W. TABELL, Editor November, 1980

-ll-

-

-

intentionally blank

-12-

-

THE OPTION MARKET -- EARLY TRENDS

BERNADETTE M. BARTELS Vice President

Shaw & Co.

-

Bernadette M. Bartels, MTA President 1977-1978, has pioneered in the collectionand technical analysis ofoptions data since the very inception of listed options trading. She has been instrumental in persuading the Options CZearing Corporation and the various exchanges to provide data which the stock market technician can utilize. She has developed a number of techniques for the analysis of options data and has been instrumental in assisting others in the development of such techniques.

Despite the expansive growth in options trading, technical analysis of options data remains in its infancy. In this article, Miss BarteZs gives an overview of the options trading mechanisms, the data currently available, and outlines some analytical techniques which have proved helpful.

THE STRUCTURE

The Chicago Board of Options began operations on April 26, 1973. It offered options on the underlying stocks of sixteen companies. It had taken four years of planning and development, but the result was a revolutionary change in option trading. Contracts with uniform strike prices and expiration dates were offered to the public for the first time. By structuring and limiting the selection of prices and expirations, offerings were concentrated, producing greater liquidity than ever before available in option trading.

-13-

To provide clearing facilities for the trading of contracts, The Options Clear- ing Corp. was organized as a Delaware corporation in 1972 under the name of Chicago Board Options Exchange Clearing Corp., a wholly owned subsidiary of the CBOE. On January 3, 1975, the American Stock Exchange purchased one half of the outstanding stock of the clearing corporation just prior to the inauguration of its own options exchange on January 13. The Philadelphia - Baltimore - Washington Exchange followed in June, 1975, the Pacific Stock Exchange in April, 1976, and finally, the Midwest Option Exchange in Decem- ber, 1976. The Midwest Option Exchange merged with CBOE earlier this year. The Options Clearing Corp. is now owned equally by the four exchanges.

The Options Clearing Corporation clears all transactions in the 283 calls and 172 puts currently traded in addition to trades in 15 dually-listed calls and 6 puts. It serves as the issuer and obliger of all listed and traded option contracts on participating exchanges. It supervises margin deposits of clearing members, i.e., cash, U . S . Treasury Bills, Letters of Credit, and shares of underlying stock. It also collects pertinent statistical data, much of which is made available to the officers and directors of each exchange. At the request of The Market Technicians Association, weekly and monthly data sheets became available to market analysts beginning the summer of 1979. Overall exchange open-interest figures are reported daily in the newspaper with the open interest in individual issues reported weekly in The Sunday New York Times and Barrons . The uncovered short position as a percentage of open sst during 1979 was estimated to average 24% by CBOE. Figures on short positions are not made public.

The Securities and Exchange Commission permitted the addition of puts to exchange option trading in June, 1977. Put options on the underlying common stock of 25 corporations were offered, in contrast to calls which, at that time, were offered on the underlying stock of 220 companies. Each exchange was allocated five puts. In October, 1977, the SEC began an evaluation study of the options market. During this period, a moratorium was imposed whereby each exchange agreed not to list options on any new underlying securities which were not listed and traded on that exchange on July 15, 1977, excep- ing the replacement of involuntarily delisted classes by new classes. The moratorium came to an end in May of this year.

There has been an increase in listings since that time. As an example, the Chicago Board of Options now offers puts on the underlying common of 50 companies versus the original 5 and calls of 120 companies versus the pre- moratorium 95. The Amex lists 78 calls and 54 puts.

REPORTED FIGURES

Public Customer

The reporting figures collected by The Options Clearing Corp. are broken down into three classifications. They are the public customer, firm proprietary , and the market maker. The term public customer includes the general public which is a combination of retail and institutional clients, but it also includes professional traders who clear their transactions through firms who

-14-

are members of The Options Clearing Corporation. All orders entered by member firms doing customer orders not marked “firm proprietary” fall into this classification. Also, a member firm does not mark on the order ticket forwarded to the clearing corporation whether the customer is retail or institutional. Without these distinctions, The Options Clearing Corp. can- not accurately refine the reporting figures within this important class.

The public customer represents a major component of overall options trading. For continuity, only figures on the Chicago Board of Options and the AMEX will be included in this article.

PUBLIC CUSTOMER

Average Monthly Call Volume

Asa %o f Total Volume

ASE CBOE

1975 57.9% N IA 1976 57.0 47.1 1977 47.0 39.4 1978 49.8 43.5 1979 50.0 40.5 1980 (8 mos.) 50.9 42.5

During the hectic bull market of 1975-1976, there was record participation by the public customer. Peak months were established on the Amex:

AMEX

Public Customer As a %

Of Total Volume

1975

May 67% June 63 July 64 1975 monthly average 57.9%

A somewhat lower record period was set in the first quarter of 1976:

-15-

PUBLIC CUSTOMER AS a %

Of Total Volume

1976

ASE CBOE

-

January February March

Monthly Avg.

61.8 53 65 55 60.3 50

vs 57% 47%

Public customer percentages tend to be more striking on the Amex due to a difference in trading structure from the CBOE. On the CBOE , a competing market system is in effect, whereas the traditional specialist system prevails on the Amex. The Amex believes its volume figures tend to be weighted on the side of the public customer.

A gradual shift in the balance of the public-customer figures away from the retail and trader influence of the 1975-1976 period seems to be taking place as institutions are accepting options as a portfolio tool. -

Peter Thayer conducted a survey of bank trust and estate departments in conjunction with his master’s thesis at Harvard University. Mr. Thayer has periodically updated the original survey, Options Achieve Respectability With Trust Departments, 115 Trusts and Estates 592. In 1973, 50% of those that responded to the survey viewed writing options as not prudent and 14% thought it too risky.

By April, 1976, the writing of options was considered neither imprudent nor risky by any of the respondents in a survey of 200 of the largest banks in the country. 81% of the banks answering the survey said they expected to be writing options within the next three years.

In the spring of 1977, all U .S . trust departments over $100 million in size, or approximately 350, were questioned. The survey indicated 76.2% of the respondents felt the Comptroller of the Currency should allow option utiii- zation beyond what was already approved.

* 29% felt trust departments of national banks should be allowed to purchase calls.

* 37% felt they should be allowed to purchase puts.

* 45% felt they should be able to write puts.

In December, 1979, the Comptroller of the Currency granted national banks the authority to establish option programs if appropriate. Mr. Thayer will soon be undertaking another update of his original survey.

-16-

.-

The Institutional Investor conducted a review of the attitude of pension funds towards options. In August, 1976, .9% of those questioned had established option programs. By February, 1979, 9% were involved in option programs with an additional 8% considering establishing programs within the year. The number of insurance companies engaged in option writing and /or’ trading is unknown.

In December, 1974, Robert Nathan Associates, in Review of Initial Trading Experience of The Chicago Board Options Exchange, estimated institutional participation to be about 5% of the total customer business. By 1976, the figure was estimated to be 10%. The CBOE currently uses a rough estimate of 12-15% which may prove to be conservative.

The growth of the institutional participation should increase the sophistication of the trading techniques within this category. The performance of this group in the market will be traced later in this article.

Firm Proprietary

Firm proprietary transactions are those for a member firm’s own account. Options may have impacted stock-exchange member trading more than any other reporting class. The growth of institutional investor activity as a percentage of stock exchange trading volume during the 1960’s and early 1970’s sorely strained the capacity of member firms and specialists to meet the liquidity needs of these large investors. Options provide a much-needed and highly effective tool. It gives member firms an alternative to the accomo- dative buy high-risk, costly process of positioning blocks of stocks. With access to options, buy side orders can be shorted and hedged by calls. Long positions can generate income through covered writing, thus reducing carrying costs. Puts can provide the block trader with time to find an interested buyer or await improved market conditions before liquidating a position. There are multiple possibilities available to firm trading departments because of options. The contribution made by options is reflected in the growth of block trading, i.e., trades of 10,000 shares or more, over the past ten years particularly since 1977.

BLOCK TRADES

9 of NYSE Volume

Year % Year 0. ro

1970 15.4 1973. 17.8 1972 18.5 1973 17.8 1974 15.6 1975 16.6

1976 18.7 1977 22.4 (option moratorium) 1978 22.9 1979 26.5 1980 29.5 (8 months)

-17-

Firm trading as a percentage of overall trading does not compare with the volume of the two other major reporting categories, public customer and market maker.

FIRM PROPRIETARY


AsA %o f Total Volume

ASE CBOE

1976 7.0 6.1 1977 7.2 6.9 1978 6.2 6.7 1979 6.5 6.6 1980 8 mos. 6.7 8 mos.6.9

The importance of the option activity by firm trading, aside from block trading, may well be in its influence on the figures member firms report to the New York Stock Exchange on their own trading. There has been a gradual increase in member-firm short sales on the NYSE as a percentage of all sales. An obvious increase in short selling participation over the last four years (1976-1979) compared to the earlier four-year period (1972-1975) can be seen from the following table. Public short figures are the third reporting category, but are not included in the table:

NYSE

Short Sales % All Reported Sales

Year

Category Short Sales Other Exchange % All Sales Specialist Members

1972 6.6 56 27 1973 6.7 53 24 1974 8.0 45 25 1975 7.7 52 28 1976 6.6 49 33 1977 7.0 42 39 1978 7.4 44 38 1979 7.1 45 38

Options may have also become an internal tool by which member firms are able to improve firm liquidity during periods of high interest rates. Im- proved liquidity lessens a firm’s dependence upon banks for expensive broker loans. Many strategies can be employed. Simply stated, a stock is sold short. A call is purchased to hedge the position. The net after- option proceeds become available for internal use, such as helping to

-

-18-

finance a customer’s margin account. As the short sale is hedged by a call, the firm’s capital reporting figure for NYSE purposes is not impacted; a customer’s margin account is financed internally; a loan at the bank is eliminated or greatly reduced in size.

Market Maker

The volume complement to the public customer is the market maker on the CBOE , and the specialist on the Amex who trades exclusively for his /her own account, making bids and offers to provide liquidity for the market. It .provides a contratrend performance in the options market to the public customer .

MARKET MAKER


AsA %o f Total Volume

ASE CBOE

1976 36.0% 46.8% 1977 45.3 53.7 1978 44.0 49.1 1979 43.4 52.9 1980 8 mos . 42.4 50.5

On the Amex there are 182 specialists registered in options of which approximately 90-100 are actively making markets in options. On the CBOE , the total membership prior to the merger with the Midwest Exchange, was 1,250, of which at least 500 were regular market makers. This is a fluid number due to the flexibility in market making provided by The Exchange. At times, as many as 700 may be acting as market makers. After the merger with the Midwest Options Exchange, the CBOE added an additional 390 special members whose trading participation is limited to the 16 issues originally traded on The Midwest. Of the speical members, 155 are registered as market makers and floor brokers. In practice, approximately two thirds of the total number are market makers.

TRADING TRENDS

Market analysts have been working with limited resources when faced with analyzing the options market and its interrelationship with the senior stock market. The Options Clearing Corp. began to issue statistics during the summer of 1979. Back data has been promised on all of the published series but remains unavailable at this time. By gradually collecting and comparing figures, certain early trends are beginning to appear. These observations are presented with some serious reservations. There is a limited amount of data available and the time frame covered is short. Conclusions reached must be classified as tentative at best.

-19-

intentionally blank

-2O-

Volume

The first is that option activity in the public customer category tends to contract as a percentage of overall volume as the senior equity market declines and conversely, increases as the senior market advances.

Put /Call Ratio

A second statistic, put activity relative to call activity, increases to ex- ceptionally high levels in all reporting categories at low points in the equity market. This spring when the moratorium was lifted, an additional 142 puts became available for trading versus the 25 traded during the prior three years. This marked increase diminished the importance of past ratio figures as a reference. Time will be required for a balance to return to this measurement as the ratio has reached a higher working plateau. Additional market swings will be needed to establish a new reference point.

-21-

.,... . . . . . .-* -@

;; -,:;j: :

1.

4 ‘1:. ;, : :;::: !’ ‘,‘. ,,_.,. I . . . .../

--I?: ,:::..I : .I:::.: . . .:,I. . . . . . . . “..‘.‘I

: :

. ,

- ,

: :

1:

1: 1..

I I . fl. . .

Asimple ratio is constructed by combining the total volume of puts divided by the total volume of calls. Weekly figures become available from The OCC in June, 1979. Since June, 1980, daily figures are reported in the newspaper.

A monthly survey of put /call activity from June, 1979June, 1980 indicated a confirmation of stock market turning points was possible through use of the ratio. In the past, the actual ratio figure had been expanding at turning points due to the limited number of available puts at a time when there was a growing demand due to increasing sophistication in option usage. Extreme levels were still easily identified.

Date Put /Call Ratio

March, 1978 16.7

November, 1978 16.2

July, 1979 17.1

October, 1979 19.0

March, 1980 18.1

Prior Month’s Ratio

11.7

12.0

15.0

15.0

11.7

Flow of Funds/Public Customer

A plethora of strategies have evolved in the options market. There are bull, bear, and neutral market strategies. To compensate for this situation, a simple netting of positions was adopted to simplify the tracing of market activity by various reporting categories. The objective is to determine whether the balance of the activity has a bullish or a bearish bias. The figures used in the following illustration are monthly volume figures of the public customer’s activity on the Amex which were collected before The OCC began to issue its statistics. Average premiums were not accumulated during this period. It is the net percentage of the public customer as a percentage of total volume. In studying the monthly activity figures of the public customer, an interesting pattern started to become apparent at stock market turning points. The public customer, on balance, tends to sell puts at low points and buy calls.

-23-

intentionally blank

-24-

It wasn’t until weekly figures became available with average premiums that the trend could be confirmed. By netting the buy/sells of the public customer, times the appropriate average call or put premium, and then accumulating the figure as one does in market breadth calculations, an interesting pattern emerges. At high points in the market, the amount of money flowing into calls drops sharply. It expands dramatically at low points in the market. A contra-trend occurs in puts. One must be cau- tioned. These figures measure only a brief period in option market history.

FLOW OF FUNDS Public Customer

PUTS

FLOW OF FUNDS Public Customer

CALLS

-25-

COtdPhRiSON'OF WEEKLY VALUES JANUARY 5, 1979 TlfROUGH SEPTEMBER 5, 1980

6-j--=+-

8

J7-M III I IV l-T-k-h 1979 1980

A sharp drop in the flow of funds into calls over a few weeks time might well be construed as warning to carefully reexamine one’s indicators of the senior stock market. A reversal of trend may be in the offing.

Premiums

Premiums are certainly an expression of investor expectations for the underlying stocks but the exchanges believe they are also influenced by actual and anticipated interest rate trends. A stock-exchange firm may be increasingly aggressive in the utilization of the options market to offset block trades and to generate cash for internal use during periods of rising money rates, creating upward pressure on premiums. Customers may increasingly take advantage of the leverage offered by options. Buy- ing options rather than the underlying stocks requires a cash commitment but helps to keep debit balances at a minimum during a period of high interest rates. The risk is partially counterbalanced by the leverage provided by lower carrying cost. When studying premiums, absolute numbers may be misleading but trends are interesting.

Both the CBOE and the Amex have developed call and put indices which concentrate on premiums but with consideration given to the expiration characteristic of options. The CBOE index was developed by Tom Rzepski and reflects premiums from that exchange. In creating the index, adjusted for the dissipating value of the option due to a set expiration date, CBOE learned premiums can decline while stocks are rising in price during a period when traders expect interest rates to drop. This occurred in 1976. In 1977 premiums reached a low seven months before the S & P (see chart, CBOE S & P 500). A detailed explanation of the methodology of constructing the index can be obtained from their Research Deoartment. Write and ask for the CBOE Call Option Index Methodology and Technical Consideration. Barrons nrints the index weeklv.

The Amex has created an industry-wide index which measures premium level of both calls and puts. The Amex index is computed daily. Des- criptive material was distributed at the Market Technician’s option meeting at The NYSSA in June. Tapes of the meeting are available from The New York Society of Security Analysts.

Averaqe Dollar Premiums

During an advancing period in the market, the demand for calls results in a higher premium being asked by the seller to write a contract. The premium provides compensation for the risk that the option will be exer- cised. In addition, the writer is capitalizing upon the profit opportunity created by market forces. As the market advances, the buyer is less inclined to pay a high premium for a put but will pay a higher premium for a call. A reverse attitude prevails during market declines. The professional option trader, individual or institution, will supply or write calls capturing higher premium income during an advancing period in the market in calls considered to be overvalued and at the same time accumulate

-27-

00s WOOd ONV CltlVONVlS i

R ln d

r(

AALLL

-77 I I I I I I I 1 1

;=

-- ---

2 -\ --

-L= 2 Y: --- --- x -=. -- -- -

J ’ ’ 1 ’ ’ ’ ’ !

m in a ln m ti .6 ci f: L6 .. d

-X3UNI NOIli -IX3 3083

,

-28

puts which are less in demand thus offered at low premiums and considered to be undervalued. Opposite procedures take place during a declining period in the market when the professional accumulates calls and offers puts.

As market forces do exert influence on premiums, there does seem to be a relevance to watching the trend of premium levels. A ratio of the dollar premiums puts/calls over the past year is outlined below :

Members of The Market Technicians Association are successfully applying the data made available to them by The Options Clearing Corp.

Gail Dudak and Dick Orr of Pershing & Co. have taken the figures provided by The OCC and constructed ratios of customer and firm buy/sell activity. It produces a measurement of option activity for the two reporting classes, but they have taken the process one step further. By producing a ratio of the ratios they have achieved interesting results. When the customer ratio/firm ratio drops below .75, a constructive attitude toward the stock market appears warranted. Note the high points.

-29-

- 30-

Firm Proprietary

Although firm figures are a relatively small percentage of overall trading, this category represents a very sophisticated market sector. -Arthur Merrill of Merrill Analysis, publishes a measurement of firm trading in his Technical Trends. It is an interesting refinement of the figures. Its objective is to -the bullish or bearish attitudes of the sophisticated trader. The formula is on the chart. It reads 100 X (Call Buy + Put Sells - Calls Sells - Put Buys). An exponential average of the figures is charted.

z +30r a 1 I I I I I 1

-,o- J,“A.,,A,O”OJ?~A~J JAAOAO

IS79 ISIO

Other Data

Barrons began to publish traditional breadth figures for the CBOE on June 21, 1980. The figures to date are:

Week calls Puts Ended Advances Declines Unchanged Advances Declines Unchanged .-.---

6120 176 923 107 316 95 100 6127 606 354 133 85 307 61 714 665 383 110 71 349 71 7111 661 433 90 108 310 86 7/18 943 187 70 43 409 60 7125 266 698 47 206 168 74 8/l 770 386 65 138 317 81 818 779 366 87 98 349 112 8115 710 461 104 87 381 113 8122 568 504 75 125 297 107 8129 231 986 62 291 178 103 915 845 337 109 106 351 121

-31-

Back data should be available on this series as the figures are collected daily. They were offered to The MTA in 1978 at which time they contained daily trading statistics from October, 1976 to August, 1978.

THE TECHNICIAN AND THE OPTIONS MARKET

The following growth pattern for one options exchange confirms a need for extra effort on the part of analysts to understand the options market. Each contract represents an option on 100 shares of stock.

CBOE

Average Daily Volume

May, 1980 - August, 1980

Time Period

May-Dee 1973 Jan. -Dee 1974 II lf 1975 II ?l 1976 11 17 1977 II II 1978 ?I . 0 1979

Jan.-Aug.1980

Avg. Daily Volume

(# Contracts)

Peak Volume Day

(# Contracts) Date

6,469 N/A 22,442 N/A 57,040 124,528 (10/14) 84,972 158,308 (09122) 98.566 169,158 (04/14)

136,021 402,440 (04117) 139,840 332,814 (lo/lo) 166,337 328,698 (01109)

The option market offers leverage and provides liquidity to the stock market. As it expands in its number of offerings and as investors’ education and appreciation of its function deepens, its growing importance will influence the entire financial industry. It is important that we as market analysts acknowledge this new force and combine our talents to study the long range implications for our specialty. By sharing our knowledge, our effectiveness as professionals will be strengthened.

-32-

SOURCES OF INFORMATION

Karen Carson, Director of Statistics, Chicago Board of Options and Beverly Gordon of the Amex. were most helpful in supplying trading statistics for the exchanges they represent.

REFERENCES

A Survey of Investors on the Listed Options Market, 1975, American Stock Exchange

Prospectus, The Options Clearing Corporation, Dated January 6, 1975, October 16, 1978.

The State of the Options Sector of the Securities Industry, May 6, 1976, Securities Industry Assn.

Options for Institutions : The Prudent Man Rule, 1978, American Stock Exchange.

Options for Institutions : Insurance Companies and Mutual Fund, 1978, American Stock Exchange.

Exchange-Traded Options on Common Stock, Federal Reserve Bank of New York Quarterly Review, Winter 1978-79, Vol. 3, No. 4 Pg. 8-26.

New York Stock Exchange, FACT BOOK 1973-1980

The Chicago Board Options Exchange - Market Statistics 1979

-33-

intentionally blank

-34-

A RANDOM WALK THROUGH RANDOM NUMBERS

BERNIE FREMERMAN

and

MORTON JACOBS, M.D.

The random walk hypothesis has been the bete noire of technicians for the past decade. Yet, in one sense, there is no such thing as randomness. Random number series are regularly generated by computer, and the programmer who writes such a random number generator will be able to predict the next number in the series with 100 percent accuracy. This phenomenon and its implications are examined in some detaiZ by Fremerman and Jacobs who choose what should be a “random” series, numbers in a Kansas City teZe- phone book, and find some surprising evidence of non-randomness.

Philosophers have observed that the human mind tends to seek order out of chaos. When we search for cycles in a time series, we can usually find them. Whether or not they are real is a good question and one which may not always have an answer. This paper is the result of an investigation into randomness and will propose more questions than answers.

The Foundation for the Study of Cycles has developed techniques to isolate periodic cycles in statistical time series. How valid are these methods? Some researchers have been troubled by a question of logic in applying them. We know that, if we assemble a synthesis of regular periodic cycles of controlled data to form a complex curve, this complex pattern can be separated into its components. This seems to imply that, if we apply the same statistical techniques to a “real” set of data and are able to separate out cyclical components, these components are real, and not caused by a fortuitous set of randoms. This is somewhat comparable to saying that if A implies B , then, given that B is true, it is implied that A is also true, which is not good logic.

-35-

-u4uooa-Yun4cn TCST

ua. OF DATA POINTS

--bATA IS ANNUAL

150 FI,,St OAlA POINI * 1401.50 BASE OAlE ‘ 1400.50

- OATC OF FlRSt 011A

-

BASK DAIC Al STAR1

_

POlNll -

“‘Nlr= 0 o.v=

5o.*u000

-. 50.11000 54.41000 ae.11000

-- 5V.13000

I 59.52OUlJ

51.13000

I' I. 5Y.69000

bJ.PL1000

50.10000

I-- 52.4lEOO

59.50000

52.11000

57.11000 --

OF STCIEC REFDRT ANY

,S.l5COO s3.s7000

52.ubLOO 51.9YOOO

59.25GOO 57.02ouu

53.2*coo 5)*.45000

55.17COO 56.40000

59;37000

. r

TABLE 1

1CLLP”ONC YUROCYS -- ____ ~__- ____ ---_. -- . -..-.- -- .- - m - -- _--.

56.YVPOO 59.1ZOUO

59.9uco; 51;9ionn

53.Y5COO 50.6*OCO

52.4rCOO sr .C"OFO

5l.b4(lOO 53.2OOOO

56.77LOO 59.53000

59.11LOO 57.6'JOCO

VI.77000 55.21)000

55.lJPoO 56.93000

59.70c00 51.26000

.- --- --.-

0 -

-

1LARX 1801

AOJUSTING IS

55.46000 55.07000

~2.35UOO 55.02000

L9;6WOO 52.77000

52.29000 52.17000

51.90c00 51.4kCOO

,,.**coo 55.Y4COO

55.5,"OO 5Y.47UOO

5l.baooo 50.64"OO

51.UlL30 5U.1~003

54.12000 5Y.Y'UOO

52.37‘"0 51).2coco

50.27000 57.91OUP

59.24000 51.24UOO

57.54000 5J.LiULIO

51.27000 56.%000

OUAHlfW= 0

1400.50

54.91000

53.09000

>t..5*000

59.69000

56.50000

51.07000

55.4"OOO

53.470JC

56.240^0

5O.dZUJO

>Y.')blJJ"

53.67CCO

59.12000

55.u7030

51.72000

54.1nooo 56.96004

54.YTOOO - 56.*6000-

5b.lbOFO 52.64000

53.15000 57.30900

50.*2000 51.42000

53.17000 5Y;l2GOO

5u.35000 52.50000

51.37000 _ 59.00000

56.6bOOO 52.69000

Bb.*Zooo 59.19000

50.9~000 5*.onooo

5S.biCOO 59.35000

52.9COOO 57.*5uoo

56.119600 54;bbOOO.

56.1~000 53.42UOo

_-.- - _. 57.74000

--57.62000

54;92000

57.07000

55.erooo

5b.70010

57.44000

5*.e3000

56.91000

56.3bOOO

52.70000

54.X050

5o.*Pooo

57;s~Ooo

59.3ioco

55.53000

5O.lJOOU

57.59000

52.*4000

56.44000

55.55100

55;05000

5*.33ooL

55.2UOOO

45.26003

57.4uooo

57.15932

55.(19000

5z.*vGoo 55.2LuOL

-36-

We recently tested some random numbers to see if they contained any significant cycles. The results were surprising and unexpected. A “time series” was assembled by selecting telephone numbers out of the alphabetical listings in the greater Kansas City telephone directory. If the last 4 digits of a number were between 5000 and 5999 inclusive, it was accepted as an item in the time series. If it was outside of this range, it was skipped. In all, 150 numbers were accumulated in the order in which they were found in the telephone directory. Each number was divided by 100 to simulate a “price” in dollars and cents. A starting date was arbitrarily selected as 1801. The result was a time series of 150 “years” of annual data. The final series is shown in Table 1 (reading in rows from left to right across the page like a typewriter). These values were converted to logarithms. The logs of the data are plotted in the graph in Chart A. They fluctuate within a narrow range and have a flat trend because only numbers between 5000 and 5999 ( common logarithms in the range 1.70 to 1.78) were used.

A Systematic Periodic Reconaissance was made of the logs of the data, from the first to the thirtieth harmonic at intervals of l/10 harmonics. This was done to get a hint of what cycles might be in the data. Chart B is a periodo- gram of the possible frequencies. As shown, a number of spectral peaks are apparent -- the strongest at the 20.3rd harmonic, corresponding to 7.38 “years. ”

The departures from a 7-term moving average were computed and arranged in a periodic table of 7.38 terms using the Foundation’s Universal Array Program. The most likely period turns out to be 7.35664 terms. The results of the Universal Array are shown in Chart C. This chart is set up in six sections. The first three are averages of each third of the data. The next two represent averages of each half of the data, and the bottom section is the average for the entire series. A sinusoid has been fitted to the data points in each section of the array. As can be seen, there seems to be a very regular and consistent cycle of 7.35 terms in this random series of telephone numbers.

-How significant is this cycle ? What is the probability that it could occur by chance? The Bartel’s test of significance was used to determine this, by comparing the sinusoids fitted to the individual waves with the sinusoid fitted to the average of all the waves. This test revealed a probability of .00028 -- which means that this cycle would have been expected to occur by chance about once in 3,546 times!

In chart D we have plotted the 7-term departures smoothed with a 3-term moving average. Also plotted in chart D is the sine curve having a period of 7.35664 terms. The relationship between the two curves is quite apparent.

What might the meaning of these findings be ? Do the statistical techniques create cycles ? Are the cycles found “real,” and what does this imply about “random” events? As to the reality of the cycle found, it would be helpful

to define more clearly what we mean by the term”rea1. I’ We can speak of statistical reality versus the concept of a real cycle as a phenomenon which exists in nature, continues into the future and so is “predictable. 1’

-37-

- if-

-38-

In this vein, to see if there is a “real” (i.e. , “continuing”) 7.35 term cycle in telephone numbers, another series of 150 telephone numbers was assembled beginning on another page of the telephone directory. The first data point was assigned the date 1951 as if it were a continuation of the first series. Four periodic cycles were found in the second set of data, each close to the 7.35 term cycle found in the first set. None had the degree of significance revealed by the Bartel’s test, but all were fairly significant. A tabulation of the results of the array analysis of the first and second sets of data is shown below:

First Set of 150 Data Points

PERIOD AMPLITUDE

7.35664 .00616

Second Set of 150 Data Points 6.82842 .01964

. 7.61132 .00520 6.87398 .00460 7.66961 .00834

PHASE BARTEL’S

TEST

1807.35 .00028 (1954.4)

1954.23 .0400 1956.09 .0570 1956.61 .0432 1955.39 .0725

In the first set of data, the 7.35 term cycle would have had a peak in 1954.4 if extended forward in time. This compares rather favorably with the peak phases for the four cycles found in the second set of data. When analyzing cycles in “real” data, we often tend to view slight phase shifts or changes in amplitude or period as some of the unexplainable features of real cycles. Obviously the results we get here bring these assumptions into question.

We went one step further and “randomly” shuffled the punched cards (each containing 8 data points) of both sets together to form a new “time series” of 300 data points. One of the strongest cycles found in this new series was 6.97 terms with a probability of .0337. whole new set of questions.

Each new approach we took raised a There are a number of other tests of “random”

numbers that we would like to make when we have the opportunity.

This entire matter sort of reminds us of a friend’s dog. trained to bring in the newspaper.

This dog has been

his dog will find one somewhere. If there is no paper in our friend’s yard,

He never returns empty-mouthed. Perhaps the techniques we use will isolate cycles in data -- if they are present. But, if there are no cycles in the data, the techniques will massage the numbers sufficiently so that the closest thing to a cycle will emerge.

-39

One of the theories of the behavior of security prices is that they follow a “random walk.” Yet we often find evidence which conflicts with this theory. We find cycles which have regular recurring patterns, and they frequently continue after they’ve been discovered. Are these patterns real, or are they the result of the unusual behavior of random forces? Does the principle of uncertainty which applies to sub-atomic particles also apply to human behavior ?

There have been some recent investigations into the very nature of randomness. In their book, The Challenge of Chance (Vintage Books, New York, N.Y., 1975)) Alister Hardy, Robert Harvie and Arthur Koestler bring this area into focus. They arrive at the conclusion that what we preceive to be randomness may in fact have definite patterns.

One of their investigations of randomness led them up a strange path. Radio- active decay is a nuclear phenomenon where it is possible to predict very precisely, the half-life of radioactive material, that is, the intensity of radiation decreases by half in a precise period of time. While this decay rate is predictable for a given mass of material, it is not possible to predict the disintegration of individual particles. The decay of single particles occurs randomly.

Helmut Schmidt, of the Mind Science Foundation in San Antonio conducted an experiment in which a set of four lamps was lit in a random sequence. The sequence was controlled by a radioactive source. Subjects who allegedly had extrasensory capabilities were to guess which of the four lamps would light up next (Or alternatively, which would not light up next. ) . In a series of 103,000 guesses, the subjects were able to guess correctly with a probability beyond one chance in ten thousand million! Does this imply that the subjects were able to effect some degree of control over these “random” events? Or does it reinforce the idea that what we perceive to be randomness is in fact not random?

Either of these concepts would appear to be borne out by our experiment. We are left with the interpretation that our 7.35 term cycle (which was regular throughout the data with only 1 chance in 3,546 of being a chance event) is the “order out of disorder” phenomenon referred to by Hardy, Harvie, and Koestler .

This still begs the question as to how, in turning to a “random” page in the phone book, a series with such great significance was formed. We feel that if each successive 150 data points in this phone book were to be studied, it is unlikely that another cycle of such regularity and significance would be found (But we wouldn’t bet on it! ) .

Was there an unseen force guiding our hand as the telephone book was opened to this page ? We are reminded of the instances of opening a book “by accident ” to the exact page in a dictionary with the desired word on it. Numerous examples of this kind of “coincidence” are given by Hardy, Harvie and Koestler as well as by many others. History is full of “unseen forces” that have affected the course of human affairs. Adam ~Smith described a

-4O-

somewhat less mystical “invisible hand” which guides the economic destiny of a free society. There are references to a “sinister force” which was allegedly responsible for the 18-minute erasure of one of the Watergate tapes.

Were we finding confirmatory evidence for an unseen force that goes beyond what is known about physical causality or were our findings the result of chance? Surely we must at least take into account the possibility of cycles in our data occurring “randomly” when interpreting the significance of our findings.

One of our friends observed that, if telephone numbers are ever actively traded on one of the commodity exchanges, we’ll have an edge over others . . .Oh well, no one ever said that the life of a prophet was easy!

-41-

intentionally blank

-42-

BUY SIDE, SELL SIDE, ALL AROUND THE TOWN

DAVID L. UPSHAW Vice President

Director of Portfolio Strategy Research Waddell & Reed, Inc.

The MTA membership roles can, by and large, be divided into two categories: Technicians who work on the “buy side” of the Street --- for money management firms who distribute commissions --- and those who work on the “seZZ side” --- for the brokers who receive those commissions. While we all consider ourselves professionals, sharing the same discipline, our point of view is natu- rally colored by the side of the Street on which we earn our living. David Upshaw, MTA’s President in 1976-1977, possesses a unique perspective. He began his career on the “buy side, ” switched for an eight-year period, and recently has returned. His articZe draws on his unique experience and offers some interesting observations about the dichotomy in our profession.

Does the Street have a sunny side ? Lf so, which side is it? (Raymond DeVoe used to ask “Is there intelligent life on Wall Street after lunch?” Ray’s is a more interesting question than mine, but the subject is beyond the scope of this article. )

In working both sides of the Street in two stints on the buy side (WaddeIl & Reed, Incorporated) and one on the sell side (Drexel Burnham Lambert, Incorporated), I think I’ve seen some of the best and worst each side has to offer. Here are some observations about my own quest for the sunny side, and, at the end, some hints to those readers currently on the sell side who want to increase their effectiveness with their buy-side clients and friends.

No question about it, sell-side technical jobs are more glamorous than buy- side technical jobs. The contrast reminds me of a WWII fighter pilot and a heavy bomber skipper. The fighter pilot had glamour, dash, and high visibility . The bomber skipper’s work was just as dangerous, but he got relatively little recognition as an individual.

-43-

The sell-side technician who gains recognition pays for it in a number of ways, including a lot of time on the road, and a never-ending phone call schedule. In return, the sell-sider meets a large number of people all over this country. A few of them are brilliant; almost all are interesting in one way or another. There is a great deal of excellent technical work being done in out-of-the-way places by people who are not well-known -- except to the sell-siders who are privileged to call on them. We need to encourage these buy-side ladies and gentlemen to join the MTA and write articles for this Journal. Otherwise, we will never know how good some of them are.

The buy-siders may be, but are not always, relatively unknown in our profession, but they, too, meet a fair number of interesting people: the sell- siders who storm their offices by personal visits, phone calls, and mail. In general, the larger (in assets under management) the buy-side firm, the more sell-side technicians its technician will know. Buy-siders are also getting to know each other better, thanks to MTA Seminars, our Journal, and other technical gatherings. In my first tour at Waddell & Reed (196819721, I met an incredible variety of sell-side technicians, including a lot of truly off- the-wall characters. In my second tour, which began in February of this year, most of the technicians I see are people I’ve known for years. There aren’t as many really peculiar people hawking technical wares in 1980 as there were in 1968, it seems. The solid types are still around, doing good work. Still, I miss the spice contributed by the people on the fringe.

Buy-side people, it seems to me, have more time to do new research work, probably because they usually have fewer people to serve than their sell- side counterparts. It is a luxury to be able to take a new idea and hound it to death until it proves to be either useful or worthless. Sell-side technicians are usually committed to a regular publishing schedule in addition to their on- going client contact work, and in my experience, it is more difficult for them to get a day or two off the treadmill to pursue new concepts. Understand that I respect sell-side work and the people who do it. “free” time is harder to come by on the sell side.

I’m only noting that

Which side is the sunny side ? That’s for each person to decide, and the deci- sion should be made, ideally, on the basis of what each person really likes to do. The sell side offers visibility, or the business equivalent of greasepaint and the roar of the crowd. It may, or may not, offer more money. (It’s about time we did another MTA salary survey .> The buy side, it seems to me, offers a somewhat more contemplative working environment and exposure to a wide variety of technicians and their work. I prefer the buy side, but I am very glad I spent eight years on the sell side. Having been on both sides, I close with some suggestions to my sell-side friends on How to Win Friends and Influence People on the Buy Side.

Realize what you are up against when it comes to getting the attention of buy- siders . My incoming mail pile ranges from four to eight inches in thickness each and every week. There is no way that I can read all that material and still do my job. If you want to get my attention, write as little as you possibly can and still get your message across.

-44-

When writing to technicians, remember that technicians are chart-oriented. What the Chinese say about pictures is true. You can tell me more with a couple of good charts than you can with thousands of words. I have the time to scan your charts; I don’t have the time to read all your text. This is a sad thing for me to report to you, for many of you are good writers and you enjoy writing. The fact is, there is more writing coming across my desk than I can cope with, no matter how good it is. The writing problem is com- pounded, I know, by the fact that you are writing not just to technicians, but to portfolio managers who may or may not be chart-oriented. All I can do here is sketch the problem, which is real. You’ll have to find your own mix of words and pictures. It’s tough. I’ve tried it.

If you have a truly momentous thought, such as that a primary bull market has begun or ended, don’t depend on the written word. Get on the phone calls; they may not read your 80-page treatise. On the other hand, don’t call just out of a sense of duty. If you have nothing new to say, don’t waste your time and your client’s time with calls that don’t have a specific purpose.

Get face-to-face with your clients at least once a year. From my sell-side days, I recall that clients would say, at the end of a personal visit, “I’ve read your stuff, but I never really understood it until you came here to explain it. Thanks for coming. ” Personal visits, as many as six or eight in a day, are tiring for the sell-side technician. The only defense of them is that they work.

If you use a computer, beware of thinking that a long printout will do your work for you. We are swamped with printouts, fundamental and technical. I use a computer, and I know that they are great tools. But a printout un- accompanied by a brief summary stressing what it is that the computer is telling you, its master, is worthless to me. I don’t have the time or patience to dig through a printout, looking for gold. The gold, if it’s there, ought to be stressed in a front-page summary.

I was guilty of all the sins I have just mentioned, and then some, when I was on the sell side. That’s why I can now point them out to anyone who cares to read about them. My comments are offered with a great deal of respect for everyone on the sell side who is trying to help me do my job. Thank you for your efforts.

- 45-

intentionally blank

- 46-

CORPORATE NEWS REPORTS AND SECURITY RETURNS

WALTER L. ECKARDT , JR. Associate Professor of Finance

Southern Illinois University at Edwardsville

and

JACK D. BECKER Assistant Professor of Quantitative Methods

University of Missouri at St. Louis

Technicians have generally agreed on the concept that fundamental news tends to be reflected in the market pZace,often before the fact. Theauthors conduct a rigorous examination of the behavior of individual stocks in response to known news events using both previously developed, rudimentary techniques of analysis and more sop his ticated ones growing out of the CAPM. All methodologies seem to produce similar results showing, among other things, the market’s tendency to react to good and bad news prior to its being reported. There remains, however, some difficulty in proving a direct relationship between specific information items and stock performance.

1. INTRODUCTION

The market impact of new information arrival has been a popular topic of investigation for students of securities prices. Proponents of the efficient markets hypothesis contend that the appearance of incremental information, itself unpredictable in terms of both timing and nature, is accompanied by an instantaneous, unbiased adjustment of affected securities values.1

-47-

A proliferation of empirical work has combined with theoretical advances to extend and refine experimental design over the years. Consequently, it is not an easy task to directly compare the results of a number of studies. This maturation of methodology raises the question of design robustness, where the burden of proof logically falls on the older investigations.

The impact of “news events” on security prices is a topic which holds considerable interest but has received comparatively little attention in the literature. In this context, a news event is defined as information which appears in a widely-read publication and which could potentially affect security values. Niederhoffer [ 111 analyzes the effects of front-page head- lines on the level of a stock market index, while Klein and Prestbo [ 91 carry out a similar investigation using the “What’s News” column of the Wall Street Journal. Reilly and Drzycimski [ 131 study the influence of sevengo world events on seven aggregate price series. The same authors [14] also relate market performance following twelve major world events to the opportunity for profitable specialist activity. The results of these studies per- taining to market efficiency are generally mixed, and efforts to obtain a detailed, valid comparision are frustrated by the inconsistency of methodologies .

Each of the investigations cited above deals with the effects of news events on aggregate market measures. With the exception of some preliminary work by McCain and Millar [ 111 , impact on individual securities has not, to our knowledge, been reported. This is not surprising, since there are at least two substantial problems associated with such an undertaking. First, the data collection and processing problems increase dramatically when a large number of individual price, dividend, and capitalization change series must be analyzed. Second, the significant positive correlation between stock price changes indicates that security-specific location-constant (i.e. single source) news should be more informative than the security-general location- constant news used by the other investigators.2 However, such a procedure exacerbates the problem of 17scoring” individual news items on a bad * good scale. As noted by Niederhoffer [ 121, reliable scoring requires evaluation replication. 3 His study involved three-man replication of 432 events, while this study utilizes four-man replication of 395 events. These necessary operations require considerable manpower and financial resources which are all too often unavailable.

The purpose of this paper is twofold. First, the impact of news events on individual common stock prices is analyzed. The information source is the widely read Wall Street Journal column, “Heard on the Street .” Second, four widely emplowechniques for assessing information effects are compared using the news events data. In this way, some indication is provided with regard to the robustness of the less formal older mgthods in relation to the theoretically superior, recently-developed schemes.

The paper is organized as follows. Section 2 provides a brief review of performance measurement methodology, while section 3 presents and discusses our empirical results. Section 4 concludes the paper.

-48

2. REVIEW OF METHODOLOGY

Information impact studies have in common the need for an economically tract- able measure of “abnormal” security performance. Three classes of measures have been widely used. Each will be briefly characterized below and representative applications will be cited. Of course, the implementation of a measure is largely a function of the specific situation under investigation. The study at hand is no exception. Accordingly, an explanation of the test structure employed here is deferred to section 3.

a) Naive Market-Adjusted Method

A commonly used measure of abnormal security performance can be expressed in the form

q it = Rit - Rmt (1)

wherev it is the abnormal return of security i during period t, Rit is the total return of security i during period t, and Rmt is the total return of a market surrogate during period t. 5 A risk control was provided by either stratifying according to some pre-specified risk class or forming portfolios with the hope of “averaging out” individual security risk. This performance measure is used by Van Horne [ 181 in a study of new listings performance, as well as Largay [ 101 and Eckardt and Rogoff [ 31 in investigations of exchange-imposed 100% margin requirements.

b) Linear Market Model Adjustment Methods

Advances in the theory of capital markets have led to the explicit inclusion of risk in abnormal performance measures. The two factor linear “market model” suggested by Sharpe [ 171 led to the following formulation

Uit = Rit - (pi +~i Rmt) (2)

where uit is the abnormal return on security i during period t, and& and pi are the estimated intercept and slope of a regression of realizatipns of

Rit against Rmt , respectively. In accordance with this model, the ,& i term, or beta coeificient , provides a partial description of the risk associated with security i. Equation (2) forms the basis for evaluation in the stock split effects research of Fama, Fisher, Jensen, and Roll [ 51, the accounting numbers influence investigation of Ball and Brown [ 11, the secondary distribution stud of Scholes [ 161, and the FRS discount rate effects results of Waud [ 191. 7

Further work by Sharpe [ 171 indicated that, under a number of assumptions, notably the existence of a riskless asset, capital market equilibrium lends to a single factor model, where

ac i = (1 - pi> RFt V i,t (3)

-49

TABLE 1

SCORING RESULTS

PANEL A: UNAGGREGATED

Score YEAR '1 2 3 4 5 TOTAL

1972 2 73 166 132 7 380

1973 '0 93 124 93 10 320

1974 5 102 121 61 11 300

TOTAL 7 268 411 286 28 1000

PANELB: 'AGGREGATED

Score YEAR 1 2 3 4 5 TOTAL

1972 0 25 29 39 2 95

1973 0 30 18 31 1 80

1974 2 29 23 18 3 75

TOTAL 2 84 70 88 6 250

-5o-

and RFt denotes the return on the riskless asset during period t. Estimating these parameters and combining (2) and (3) yields

U*it = (Rit - RFt) - pi (Rmt - RFt) (4)

As a practical matter, this formulation has important implications for researchers who do not possess the basic data necessary to construct,,historical linear regressions relating realizations of Rit and Rmt. Values of fl i are available for a large number of stocks and RFt can be readily proxied by using Treasury bill yields. The price paid for the simplified requirements of (4) is acceptance of the assumptions that led to (3). 8

c) CAPM Specification Method

Initial empirical discrepancies between the CAPM as originally specified and actual experience led to the development of a number of alternative models. In an effort to test contending theories, Fama and MacBeth [6] suggested that security returns are generated by the following stochastic process

ii it = ?Ot + ? It + Z’it (5) P

where y Ct and glt are market-determined random variables which fix the relationship between return and risk and z ’ it is a stochastic residual term. This process is consistent with the two leading capital asset pricing formulations and it has been shown to accurately describe historical returns. 9 Equation (5) can be cast into the testable form

t it = Rit (6)

where C it is the abnormal performance value for stock i_during period t , and 3 *C period t. i 0

and y *It are estimates of the realization of Y Ct and 2 .lt for This procedure has been used by Jaffe [ 7,8] to describe returns

to corporate insiders.

3. EMPIRICAL RESULTS

Following Niederhoffer [ 121, we used a five-point subjective scoring system with the categories 1 = very bad, 2 = bad, 3 = neutral, 4 = good, and 5 = very good. Of the 395 scored events, necessary information for subsequent processing was available for 250. What follows employed this subsample. Table 1 summarizes the scoring results. Panel A indicates the unaggregated scoring distribution (four observations per event), while panel B contains one representative score for each event .11 The contents of panel B are used in subsequent computations. The observations are distributed evenly throughout the 36month sampling period (l/72 - 12/74). Not surprisingly, the dominant theme shifts from good news in 1972 to a balanced position in 1973, and finally to bad news in 1974. Extreme scores (1 and 5) are quite uncommon on an individual basis (panel A) and become scarcer yet after aggregation (panel B). Accordingly, scores 1 and 2 are henceforth merged into the category “bad news, ” while scores 4 and 5 are likewise combined to form “good news. I’

-51-

Equations (1)) (2)) (4)) and (6) were used to compute a series of abnormal performance residuals corresponding to each event. The residuals are de- noted naive, Sharpe, one-factor, and Fama-MacBeth, respectively. l2 A 22- day interval surrounding each event date (10 days before through 12 days after) was selected and eleven residuals were generated using prices on days -10, -5, -2, -1, 0, 1, 2, 3, 4, 5, 8, and 12, where the event date is day 0. Figures l-4 display cumulative average residuals for good , neutral, and bad news using each performance measure. 13

There exist a number of problems associated with the foregoing tests. In the area of experimental design, we assumed that the realizable stock price on the event date was the high, close, or low for good, neutral, or bad news respectively. While reasonable, this procedure could induce an unjustified, “overreaction/snapback” effect around the event date. We used the Value Line betas of 12/31/74 in the one-factor model, which may introduce some bias, the seriousness of which is a function of the temporal instability of beta and the estimation procedure employed.

There are statistical difficulties as well. The existence of contemporaneous or closely spaced events introduces the possibility of cross-sectional correlation between the residuals. In addition, the well known errors-in-variables problem associated with the estimation of Y tit and f *it leads to bias in the Fama-MacBeth residuals.

The most disturbing considerations, however, are theoretical. First, the CAPM is a single period equilibrium model. We used monthly observations to estimate betas that were subsequently employed to construct residuals corresponding to trading intervals ranging in size from one to five days. In addition, since the Fama-MacBeth [61 g *Ct and y *lt estimates were formed with monthly information and are intended to proxy returns, we linearized these values to correspond to the differencing interval at hand. Although ex-post bias was avoided by using the last available values, the estimates we employed implicitly assume that the returns processes repre- sented by the $ *itIs evolve smoothly over the monthly differencing interval.

The foregoing obstacles notwithstanding, Figures l- 4 appear quite similar. As expected, pre-event performance of good news events is positive and bad news is negative. Neutral news exhibits slightly negative pre-event behavior. Aside from a (possibly exaggerated) “snap-back” on day 1 for good news and bad news, post-event performance in inconclusive.

To focus more sharply on pre- and post-event behavior, we computed the residuals -10/O and O/12 for good news and bad news events. After hope- ful consideration of the convergence property of reasonably well -behaved random variables, we constructed t-tests (against zero) for the pre- and post-event average residuals. The results are presented in Table 2 (good news) and Table 3 (bad news). l4 The first line of each table entry corresponds to the -10/O residual, while the second line represents the O/12 measure. Despite the substantial differences in the four performance measures coupled with the design, statistical, and theoretical difficulties discussed previously, the results corresponding to each measure display

-52-

FIGURE 1

4

3

2

I

0

-I

-2

-3

-9

-5

-c

-7

-I

-9

I-- b-

-\

U goodnews A neutralnews

0 badnews

-53-

I

3

2

1

0

-I

-1

-3

-4

-5 -L

-7 -0

-9

FIffTIIE 2

~IVE AvERA(;E REZ33DUALS-- SI-LF1RpE c..

I

-. rh =n

IQ goodnews

A neutral news

l badnews

-54-

FIGURE3 -

l goodnews

A neutralnews

l badnews

-55-

I

FIGURE 4

CUMULATIVEAVERAGERESUXIALS-FAMA-mm

-5 4 -7 -8 -4 I

-10

I goalnews

A neutral news

0 badnews

-56-

CATEGORY #ooss

All 94

+2 41

1973 32

1974 21

TABLE 2 BEFORE/AFTER AVERAGE RESIDUALS

GOOD NEW

ave t -

.031 3.22*

-.017 -2.64*

.033 :2.38*

-.031 -3.33*

.017 1.19

-.005 - .43

.046 1.95

-.009 - .71

* significant at 5X level.

/slDuALs\ NAIVE SHARPE

ave t

.027 2.77*

-.020 -2.90*

.029 1.98

-.032 -3.41*

.012 .81

-.006 - .46

.046 1.93

-.018 -1.30

ONE-FACTOR ave t

.031 3.31*

-.016 -2.49*

.033 2.36*

-.030 -3.22*

.018 1.26

-.003 - .28

.048 2.06

-.009 - .79

ave t -

.033 3.09*

-.018 -2.18*

.035 2.43*

-.030 -3.14* I

.018 1.04 !

.002 .ll

.049 1.90

-.022 -1.16 1

-57-

CATEGORY #OBS

All 86

1972 25

1973 30

1974 31

TABLE 3 BEFORE/AFTER AVERAGE RESIdUALS

BAD NEWS

ave t -

-.087 -8.83"

.017 1.78

-.075 -7.04*

-006 .55

-.083 -7.26*

.015 1.13

-.lOO -4.31*

.029 1.32

SHARPE ave t -

-.086 -8.89*

.020 2.04*

-.076 -7.33*

-007 .61

-.081 -7.08*

.017 1.37

-.098 -4.32*

.033 1.50

ONE-FACTOR ave t -

-.086 -8.31*

.020 1.94

-.076 -6.49*

.006 .50

-.088 -7.29*

.015 1.19

-.092 -3.78*

.037 1.54

FAMA-MACBETH ave t -

-.090 -7.74*

,015 1.29

-.070 -5.59"

-.002 - .14

-.081 -6.03*

,025 1.54

-.115 -4.24*

-020 .74

* Significant at 5% level.

-58-

striking similarity. Average residuals and t values are in close agreement throughout. Tables 2 and 3 substantiate the visual interpretation of Figures l-4, although significant post-event performance for good news in 1972 led to borderline (at 5%) significant (negative) nest-event nerformance for the entire good news samnle. 15

4. CONCLUSION

This study of the impact of news events on individual securities returns is consistent with two diametrically opposite sets of conclusions. On the one hand, the apparent robustness of the abnormal return alternatives supports the contention that investigations employing conceptually unsophisticated techniques somehow lead to results that are in agreement with those obtained using sounder methodology. This heartening development allows the aggregation of evidence provided by a large number of studies concerned with the impact of incremental information on securities returns.

On the other hand,however, the inability of the theoretically superior techniques to distinguish themselves leads to the contention that the information effects we are attempting to examine in the context of a specific returns- generating process and market equilibrium model may not be possible to isolate. This conclusion has some intuitive appeal. For short-run effects such as those presented in this paper, the market moves comparatively little most of the time. Market performance adjustments like the four employed here amount to little more than a location shift, especially when average values of relatively large samples are computed. As the observation period length- ens, superior measurement techniques may well have an opportunity to mani- fest themselves, but the link between a specific information item and security performance becomes more tenuous as the distance from the event increases.

we wish to thank our four scorers -- David Huber, Leroy Chasteen, Len Suess, and Louis Hermann -- for their painstaking efforts. Also, we thank John Morris for programming and implementation assistance.

-59-

FOOTNOTES

1. Fama [ 41 contains an elaboration of the efficient markets position along with a review of related empirical work.

2. The constant-location requirement is utilized in an attempt to focus on the incremental impact of information appearing in a single, widely available news source. This eliminates the necessity of searching for the earliest public appearance of a piece of news, a laborious procedure which, even if successful, requires a judgment of access availability and timing. Of course, we have chosen an information source believed to be (on balance) timely, since the incremental effect of a repetition of widely disseminated information is likely to be negligible.

3. The other cited news events studies [ 9, 11, 12, 13, 141 employ a small number of events, each of which could be “obviously” scored.

4. We proceed with an awareness of the implications of Roll[ 141 concerning the verifiability of the CAPM.

5. Total return is expressed in the familiar form, Rit = (Pit + Dit)/(Pi, t-l -l), where Pi, t-l and Pit are the share prices of security i at the end of periods t-l and t, respectively, and Dit is the cash dividends per share paid on security i during period t.

6. Details and implications of capital asset pricing theory are tangential to the purpose of this paper, and, for the most part, are not discussed. Fama [ 41 provides a great deal of this material as well as substantial further references.

7. Actually, [ 51 and [ 191 deal with logarithmic returns, while [ 11 and [ 161 use return directly. However, it is explicitly noted in [ 11 , [ 53 , and [ 161 that both formulations yielded substantially identical results.

8. It should be stressed that the CAPM is concerned with expected returns. The market model is one particular stochastic returns generator that is consistent with the specifications of the CAPM when its parameters are suitably defined. Whether it is consistent with reality is quite another matter!

9. Test results can be found in Fama and MacBeth [ 61 and Black, Jensen, and &holes [ 21.

10. The estimation procedure is given in Fama and MacBeth [ 61. Their values will be used in this study.

11. The aggregation algorithm proceeded as follows. If all (four) scores are 3 or if the scores straddle 3, an aggregated score of 3 is assigned. If the foregoing conditions are not met, the score other than 3 that appears most often is chosen, with ties going to the most extreme score (away from 3) provided the extreme score was specified at least twice.

-6O-

12. The 2 ‘s and p” ‘s for the two-factor model were generated from the CRSP monthly price relative file. Sixty observations were input to the OLS estimation, where the last data point corresponds to the month preceeding the event date. The same p ‘s were used in the Fama-MacBeth computations.

13. Cumulation and averaging are arithmetic.

14. We computed similar statistics for various cross-classifications of the date (e.g., by month, day of week, type of news). No further insights were forthcoming and the analysis was hampered by small sample size. Details are available upon request.

15. We conducted a final test, computing the Sharpe -10/O and O/12 residuals using the 60-month betas and alphas described above compared with 25-year (300 month) betas and alphas estimated using the CRSP data file. The reduced samples of 82 (good news) and 73 (bad news) yielded nearly identical results, which are not presented here.

-61-

REFERENCES

1. Ball, Ray and Brown, Philip. “An Evaluation of Accounting Income Numbers, ” Journal of Accounting Research, 6 (Autumn, 1968) 159-78.

2. Black, Fischer, Jensen, Michael C . , and Scholes, Myron. “The Capital Asset Pricing Model : Some Empirical Tests. ” Studies in the Theory of Capital Markets. Edited by Michael C . Jensen. New York: Praeger, 1972.

3. Eckardt , Walter L. , Jr. and Rogoff, Donald L. “100% Margins Revisited ,” Journal of Finance, 31 (June, 1976) , 995- 1000.

4. Fama, Eugene F. “Efficient Capital Markets: A Review of Theoretical and Empirical Work, ” Journal of Finance, 25 (May, 1970)) 383-417.

5. Fama, Eugene F. Fisher, Lawrence, Jensen, Michael C. , and Roll, Richard. “The Adjustment of Stock Prices to New Information,” International Economic Review, 10 (February, 1969)) l-21.

6. Fama, Eugene F. and MacBeth, James D . “Risk, Return, and Equili- brium: Empirical Tests, ” Journal of Political Economy, 81 (May /June, 1973)) 607- 36.

7. Jaffe, Jeffrey F . “Special Information and Insider Trading,” Journal of Business, 47 (July, 1974)) 410-28.

8. ---------------- “The Effect of Regulation Changes on Insider Trans- actions,” Bell Journal of Economics and Management Science, 5 (Spring, 1974)) 93-121.

9. Klein, Frederick, C . and Prestbo, James A. News and the Market. New York: Dow Jones and Company, 1974.

10. Largay, James A. “100% Margins: Combatting Speculation in Individual Issues, ” Journal of Finance, 28 (September, 1973)) 973-86.

11. McCain, John E. and Millar, James A. “Price Effect of Public Analysis of Common Stocks: A Preliminary Result. Abridgment. Journal of Economics. (Missouri Valley Economic Association) , (1976) , 127.

12. Niederhoffer , Victor. “The Analysis of World Events and Stock Prices,” Journal of Business, 44 (April, 1971)) 193-219.

13. Reilly, Frank K. and Drzycimski, Eugene F. “Tests of Stock Market Efficiency Following Major Events, ” Journal of Business Research, 1 (Summer, 1973)) 57-72.

14. ------------------ “The Stock Exchange Specialist and the Market Impact of Major World Events,” Financial Analysts Journal, 31 (July- August, 1975)) 27-32.

-62-

References (Continued)

15. Roll, Richard. “A Critique of the Asset Pricing Theory’s Tests-I,” Journal of Financial Economics, 4 (May, 1977)) 129-76.

16. Scholes , Myron S . “The Market for Securities: Substitution Versus Price Pressure and the Effects of Information on Share Prices,” Journal of Business, 45 (April, 1972) 179-213.

17. Sharpe, William F. “Capital Asset Prices: A Theory of Market Equil- brium Under Conditions of Risk,” Journal of Finance, 19 (September, 19641, 425-42.

18. Van Horne, James C . “New Listings and their Price Behavior ,I’ Journal of Finance, 25 (September, 1970)) 783-94.

19. Waud, Roger N. “Public Interpretation of Federal Reserve Discount Rate Changes : Evidence on the ‘Announcement Effect’ ,” Econometrica, 38 (March, 19701, 231-50.

-63-

-

intentionally blank

-64-

THE SHORT-TERM TRADING (ARMS> INDEX REVISITED

JOHN R. MCGINLEY , JR. Van Cleef, Jordan and Wood, Inc.

The Short-term Trading Index (the advance/decline ratio divided by the up volume/down voZume ratio) is probabZy the most widely anaZy- zed of any of the modern stock market indicators. Its inventor, Dick Arms, has twice spoken before the MTA , and various members have done extensive work on the subject, including your editor who discoursed for an hour at the 1979 seminar on an extensive computer study thereof. John McGinley, the author of this article, has also written previous pieces on the subject for this pubIication and for the Newsletter.

Nonetheless, the article demonstrates that, if one is willing to spend the time, it is always possible to come up with something new. John expZains his simplified calculation of a logarithmic moving average and discusses results achieved by plotting the data on a point-and- figure basis.

In the November 1978 Journal, I wrote about the use of logs when making moving averages of the Short-term Trading Index (STI). I have since done more work in the techniques I described therein and have incorporated some ideas from Dick Arms, the inventor of the STI. Charts of the results are attached to this article.

The neutral position of the ST1 is unity because it is a ratio. This is well known. Not so well known is the mathematical oddity of ratios: there is as much room under unity (one) as there is above it, i.e., in ratioland, the distance from one to zero is the same as the distance from one to infinity! In stock market terms, twice as much volume in the average up stock as compared with the average down stock - ST1 0.50 - is the exact opposite of ST1 2.00, or twice as much volume in the average down stock. But unless you use logs, the average of these two numbers (0.50 and 2.00) is 1.25 when it should be one.

-65-

. . - _- - . . - . - . - . . . . ~_

-_ ._ ___‘. I 1

SHORT -TERM TRADING INDEX

lo-Day ModMA of the logs of the Index

: JOHN R. MCGINLEY. JR. .__- . : .-. :...Lt-p i. . ! . :-L- . -. ..-:

i,ro _ _ _. . . . . . . .--I-~.~-- . (. . .~.-1 .-... L. .-. .-. -_-- . .~_ ___ ~._.__ .._ -. -~.c.

. .-... . .._-- ; _ i

-66-

_-

-67-

The solution, as I mentioned, is to use the log(of the STI) every day in your computations. Add up the last, say, 10 day’s logs, divide by ten and take the anti-log (turn it back into a number). In this manner, one day’s closing of 2.68 won’t throw your moving average into a cocked hat, won’t move it too sharply into bearish territory, won’t turn you prematurely bearish - or grey!

Being basically lazy, I take this technique one step further and use the ModMa which I wrote up in the Newsletter a year or two ago. (For a copy, please feel free to contact me. Soon to be a major motion. . . > . In a nut- shell, instead of using all ten pieces of back data, I simply take yesterday’s MA, subtract it from today’s STI, divide this by the number of days in the average, and add the answer -- keeping the sign - to yesterday’s MA. For example, yesterday’s MA is 0.90, today’s ST1 is 0.70 and the length of the MA is to be ten days: 0. ‘70 minus 0.90 and divided by ten is minus 0.02. Minus 0.02 added to 0.90 is 0.88, the new ModMA. I thereby simplify the math, lose the large drop-off problem, and get a smoother “average .”

It has since been pointed out to me that the number you divide by, the length of your MA, in this type of computation has the effect of a regular MA of double the length, e . g. , dividing by ten in the ModMA makes it more like a 20-day normal MA. This struck a chord with me, since Dick Arms, when he spoke before the Market Technicians Association last winter, mentioned that his work showed a 21-day moving average to have above-average forecasting ability!

Thus encouraged, I decided to see what my lo-day ModMA looked like going back as far as I have been calculating in this manner, i.e., back to early 1978. An article earlier this year, in Investor’s Intelligence, gave me the idea of attempting to plot the data on a point-and-figure basis, rather than in conventional time-chart form. My first effort, a 3X affair for the year 1980 seemed to contain too much noise. So I began anew at 5X. The reader might want to study the charts for a moment before continuing.

Much more work probably needs be done, but eyeballing the longer chart, a number I had empirically noted begins to emerge, i.e. , 0.94. Time and again it is the turning point for moves, up and down. Thus initially, I suggest becoming bearish when the ModMA (of the ST1 logs) gets to 0.95 or above, and becoming bullish when it drops to 0.93 or below. * I leave it to the reader to check the dates of these turns (marked on the chart) with his own work. Because only 2+ years is such a short time span, I hope this work will encourage those with the facilities to push the log ModMA (or log MA) back in history.

*Just why this turning point is in ” bullish” question.

territory is an intriguing Possibly the bullish bias to the market in general, since the

30’s, can explain it. Maybe it is the long-sought proof that one of the attributes (burdens) of being human is being basically bullish. . .

-68-

ANOTHER POINT ABOUT POINT-AND-FIGURE

DAVID L. UPSHAW Vice President

Director of Portfolio Strategy Research Waddell & Reed, Inc.

Pointand-figure-charts have Zong had their adherents. However, few technicians have been terribly creative in creating variations on the standard theme of one-point unit charts on individual stocks. One such possible variation, long-term charts of group averages, based on monthly average prices, is discussed here together with some of its unique advantages and limitations.

Point-and-figure charts can show a long price history on a relatively small piece of paper. Compared with daily bar charts, they are easy to maintain. Aficionados post them using a wide number of units, usually ranging from $-point to 3 or 5 points, depending on the price of the security and the trading detail desired. All in all, point-and-figure charts are versatile tools.

This article illustrates point-and-figure technique applied to The Really Big Picture. Three charts are shown. They are conventional one-unit, one- point reversal charts of the S & P 400 from 1970, the S & P Aerospace Index from 1970, and the S & P Drug Index from 1970. All are posted through August, 1980. They are conventional charts except for one thing: they were constructed using only monthly average prices as published in Standard & Poor’s Securitv Price Index Record and in S & P’s The Outlook.

These charts have all the virtues mentioned at the beginning of this piece: compactness (ten years of the S & P 400 occupies 30 columns), ease of main- tenance (only one posting each month, and in some months, no posting at all), and a variety of possible posting units. In addition, I think that they show The Really Big Picture with great clarity. Readers can judge that for themselves..

-69-

I50

140 I

/ 130

1

,

I

, Ix)

I

I

j 110

\

loo

90

80

70

-_ --.L . :.::; .., ..+, : . . . . . . i-..+ .,... ..‘.. +err. . . . ::..

j I

::. . .*A

,++.L ,..* .I.. .‘T. +,i, . . . . -..- +.+* . . . . . .

;:. .,

i’,+ I(.

- _ I..

160

II0

100 I

260

250

240

230

220

210

190

180

170

160

150

140

130

./ , -.

-.

.‘.i

i 1:

W DRUGS / ‘- ---. i ..

. . ..I c---i-.- _..._ -t .‘ I

.-

.- ..i .._~.

-72-

The job I now have requires me to look at group action more than the behavior of individual issues. I drew these three charts as an experiment, and I plan to draw others. Each of my scratch charts took only about 20 minutes to construct; the charts you see were dressed up for publication.

Are you a wave theorist? Aerospace, beginning in 1974, shows, by my reckoning, three massive and almost symmetrical upward waves. Drugs went through three upward waves, suffered three or four downward waves, depending on how one counts them, and now appear to be in upward wave number 2 of a new cycle.

Do you make projections based on counts ? I caution you about this because the charts only have maximum of 12 price plots a year. All the other X marks merely connect the 12 monthly figures. I doubt if conventional count- ing will work, but, as long as you understand the limitations of the charts, you are welcome to try any technique you now use.

Do you like to draw trend lines ? These charts are as amenable to a pencil and ruler as any I have seen.

Once in a while an old indicator, or technique, can take on a new meaning if it is looked at from a different perspective. Good tools never really wear out; their users wear out because they get into ruts or develop tunnel vision. Look through your own collection, and I’ll bet you will find a few old friends that can serve you well in ways you didn’t intend when you started using them.

-73-

-

intentionally blank

-74-

CYCLES DICK A. STOKEN Published by McGraw-Hill Book Company,

New York

Book reviews

-

Reviewed by William DiIanni, V.P . Wellington Management Co.

-

_-.

In this truly excellent book,Dick Stoken attempts to explain the nature and meaning of business and stockmarket cycles, and how to better profit by them.

It is not a technical book in the vein of traditional books on subjects regard- ing market analysis. Here the author provides an in-depth understanding of how the investment world really works. The anatomy of a cycle and the broad implications of its inherent mass psychology are important inputs to the understanding of the long swings in the economy and stock movements.

Later in the book he shows how buy and sell decisions can be made with more confidence and how to identify market leaders early.

Mr. Stoken painstakingly portrays the long overriding wave similar to Kondratieff and clearly gives many of the reasons why it exists and is still with us. He further breaks down smaller segments of other cycles which are more important to one’s own investment lifetime. These cycles are broken down further to include intermediate cycles. As he says:

“During each long economic cycle there appear to be five intermediate cycles. (A three-staged expansionary phase followed by a three-staged depressionary phase has occurred during each long cycle, giving us five intermediate cycles. > An intermediate cycle consists of a period of sus- tained economic growth -- that is, a series of traditional business expansions (the two-to-four-year variety) broken by only mild recessions -- followed by a serious contraction. ”

He then details such a cycle from 1921 to 1933, (and all others). Insights into the way psychology reacts to mild contractions and the resulting will- ingness of investors and businessmen to increase risk-taking as opposed to risk-aversion are useful to understanding the meaning of the accelerating phase of successive intermediate cycles. At last -- the inevitable top and its ensuing serious contractions are mercilessly explained. . . as the after- math of 1929. All cycles to the present are similarly disected with many graphs and tables itemizing each stage.

_- -75-

This book was published in 1978, and understanding the past is much easier than reading the future. But the author attempts a forecast based on his analysis and understanding of cycles.

“The evidence suggests that we are well into the fourth long- term economic expansion of the industrial revolution. In this case, the next five to seven years are likely to see a less activist society; a fading of the problems which plagued us in the late sixties and early seventies ; and a return of political stability together with a fiscal and social conservatism. And, most important, this is likely to be a time of noninflationary economic growth -- the next recession should be mild -- accompanied by a broad bull market in stocks. The Dow Jones Industrial Average, if it is to equal the rise of 1800% which occurred during the last expansionary period, should sell at 2700. On the other hand, a fairly conservative objective, say a rise one-half as much as that in the prior expansionary period or a 900% rise, would still see the DJIA reach 1600.

“However, these years of prosperity are likely to be built on a foundation of sand. The serious recession of 1973-1974 did not. . .etc.”

You will have to read the book for explanations and details of his glimpse of the future.

Mr. Stoken has spent eight years in intense research on this book; the results are well worth his effort. While he does not mention Dow Theory or Elliott Wave Analysis, students of both or either theory can readily see how all are inextricably interwoven corroborating each other. In fact, each adds to a higher understanding of the other.

No serious market student can say his knowledge is complete without reading this book as many times as it takes in order to acquire a complete working knowledge of the subject. If you think one reading is enough to grasp this material, you are either very bright, or you’ve completely under- estimated the complexity of the matter.

-76- -.

Journal of Technical Analysis (JOTA). Issue 09 (1980, November)

Economy & Finance