The "Small Company Effect" - that is, the tendency
of "smaller" listed companies to outperform the
"market" - was first reported in 1978 by Rolf Banz.
Banz collated data on stocks listed on the New York Stock Exchange from 1931 to 1974. In each of those 43 years he divided the stockmarket into five portfolios based upon the market capitalisation of each company (i.e. the first portfolio "owned" the largest 20% of all listed shares, the last portfolio owned the 20% of companies that were the smallest).
Over the 43 years of this test, Banz found that the portfolio that held the largest companies under-performed the market by 1.3% per year, while the "smallest" companies outperformed the market by 5.5%.
In 1982, Professors Thomas Cook and Michael Rozeff repeated that testing on over 3000 stocks listed on the NYSE, AMEX and "over the counter" markets between 1968 and 1978. They divided shares into ten groups (based upon each company's stockmarket capitalisation) - and discovered similar results: The largest 10% of companies under-performed by 4.2% per year, while the 10% of "smallest" shares outperformed by 5.4% per year.
James O'Shaughnessy's recent work (for his book "What Works on Wall Street") found similar results for the 43 year period from 1951-1994. O'Shaughnessy found that "large stocks" (i.e. approximately the largest 10% of companies) and "mid-cap" stocks (i.e. approximately the second largest 10% of companies) under-performed by about 2.7% per year while "micro-cap" stocks (i.e. with capitalisations below US$25 million, or approximately the "smallest" 30% of listed shares) outperformed the market by 10.4% per year!
In "Stocks for the Long Run", Professor Jeremy Siegel writes that the small company effect "is positive in every country where it has been tested and quite significant in most of them". However he also notes that this effect "waxes and wanes" over time. For example, much of the excess performance of smaller companies in the US occurred between 1975 and 1983 (when these shares "boomed") and "smaller" shares can involve higher transaction costs (owing to a wider spread between the bid and offer prices quoted).
Investment Implications: Younger investors (who have a long term investment "horizon" and who can afford to take some extra risk) should invest part of their portfolio in some of the "smallest" companies listed on the sharemarket. Older investors (seeking to minimise risk) should aim to invest in shares below the top 10-20% by size as these offer better returns than the very largest listed companies.
The first study of "neglected", or
"unpopular", shares was published in 1964 by Professor
Between 1954 and 1961 he constructed a portfolio of 30 "popular" stocks (being the most widely owned stocks from a survey of 80 large mutual funds) and a "less popular" portfolio of stocks held by only one or two of these funds.
Over the eight year period the "popular" portfolio under-performed the market by 2.7% while the "less popular" portfolio outperformed by 0.9%.
In 1982, Professor Avner Arbell and Paul Strebel published the results of their study of 500 NYSE listed companies for five years from 1972-1976. They divided these shares into three (approximately equal) groups based upon the number of sharebrokers' analysts preparing profit forecasts.
The group of stocks that was most widely researched was found to have under-performed the market by 4.6%, while the group of least researched stocks outperformed by 6.5%.
This sample was also broken down by company size to see if the "small company effect" was causing these results. That is, to see if the least researched stocks were of "smaller" companies and if the "smaller company effect" was producing these results.
This showed that "neglect" was dominant over the "small company effect" but also that "neglected" and "smaller" companies yielded the highest investment returns (i.e. the "least researched" among the "smallest" 50% of stocks yielded the highest returns).
A year later these Professors published another study of 510 NYSE, AMEX and "over the counter" stocks over a ten-year period from 1971-1980. Returns were measured based upon market capitalisations and institutional ownership (i.e. the stocks were divided into three groups, with the "neglected" stocks held by only one institution or by none).
The most "widely owned" stocks under-performed by 5.8% per year and the "neglected" stocks outperformed the market by 5.6% per year.
Splitting the results by size showed that "neglect" dominated the "small firm effect" - suggesting that the "small firm effect" may be caused by "neglect".
Investment Implications: All research into stocks "neglected" by sharebrokers and/or "neglected" by institutions shows superior returns for "neglected" shares and inferior returns by "widely followed" and "widely owned" shares.
All investors should therefore seek to own shares that are "neglected" by brokers and have few (or no) institutional shareholders - while avoiding companies "followed" by many brokers and where many institutions already hold significant shareholdings.
We are not aware of any published research that proves that a
large shareholding by management is "good" for the
company's investment performance.
However, this idea is intuitively attractive. If management have a large stake in the company, then their interests will co-incide with those of the public minority shareholders.
In addition, the very smallest listed companies tend to have large management shareholdings (and the very largest companies have a negligible percentage of their capital held by management). The positive impact of a large management shareholding may therefore be the cause of the superior returns earned by "smaller" companies (i.e. the "small company effect).
Investment Implications: While we cannot quantify the importance of a large management shareholding, we would rather invest in a company where the CEO's financial interest is a million dollar (or ten million dollar) shareholding than a company where the CEO's financial interest is limited to a million dollar salary package.
One of the first studies of the Price/Earnings ratio and
investment returns was published in 1960 by Francis Nicholson.
This study covered 100 large stocks in each of four periods of
five-years (i.e. twenty years in total). Stocks were ranked by
P/E ratio and divided into five portfolios. Overall the highest
P/E ratio portfolio (i.e. the most "over-valued")
under-performed the market by 1.8% per year while the lowest P/E
ratio portfolio (i.e. the most "under-valued")
outperformed by 4.7% per year.
A study published in 1977 by Professor Sanjoy Basu, covering 1400 stocks for fifteen years from 1956-1971, yielded almost identical results with the high P/E ratio portfolio under-performing by 2.8% per year and the lowest P/E portfolio outperforming by 4.2% per year.
Later work examining these results broken down by company size revealed that (1) high P/E shares under-performed regardless of company size and (2) "small" companies with low P/E ratios outperformed the market very strongly.
This result is contradicted in the recent study by James O'Shaughnessy (for the 43 year period from 1951-1994 mentioned previously) which found that a portfolio of the fifty highest P/E stocks - selected from the whole market - underperformed by 4.0% per year, but that the portfolio of fifty stocks with the lowest P/E ratios also under-performed by 1.3%.
The P/E ratio only "worked" successfully when applied to "larger" companies. Here the portfolio of the highest fifty P/E shares under-performed by 2.0% while the portfolio of the fifty lowest P/E shares outperformed by 1.9%.
O'Shaughnessy's study suggests that the Dividend Yield is also a valuable selection criteria but only when applied to the "larger" companies. This is similar to Michael O'Higgins method (published in "Beating The Dow") of buying the ten highest yielding stocks (or alternatively buying the five lowest priced of these ten highest yield stocks) from the Dow Jones Average of 30 stocks of large companies.
A high yield usually indicates a low share price (as a company is "out of favour" or experiencing some "problems"). Large - and financially strong - companies can survive these "problems" (so are good investments), whereas a "smaller" company experiencing "problems" may well fail.
O'Shaughnessy's study suggests that the Price/Sales ratio is the most reliable "fundamental" statistic. His low P/S ratio portfolio selected from the whole market outperformed by 3.0% per year, while the high P/S ratio portfolio under-performed by an extremely significant 8.3%!!!
Applied to only "larger" stocks, the low P/S ratio portfolio outperformed by 2.3% per year, while the high P/S ratio portfolio under-performed by 2.1% per year.
Investment Implications: There are some contradictory results relating to "smaller" companies trading at low Price/Earnings ratios and high Dividend Yields, but the Price/Sales ratio appears to be a very useful statistic - for both "larger" and "smaller" companies.
All investors should seek low P/S, low P/E and high Yielding shares, while avoiding high P/S, high P/E and low Yielding shares.
Buying by "insiders" (i.e. directors and senior
management) and Share Re-purchases (i.e. where a company buys
back its own shares on the market) are widely considered to be
favourable. Knowledgeable "insiders" are the best
placed to know what a share is really worth.
An early study by Professor Shannon Pratt and Charles DeVere monitored 52,000 "insider" trades in 800 NYSE stocks in the seven years from 1960 to 1966. A "buy" signal was considered to have occurred when three "insiders" bought shares within one month, while three sellers within a month constituted a "sell" signal.
Stocks with "insider" buying were found to outperform shares with "insider" selling for up to three years after the "insider" transactions. The "buy" group had risen an average of 59.1%, while the "sell" group was up only 27.1%. The buy group steadily outperformed the sell group throughout the first 24 months after the "insider" signals - with both groups showing approximately similar rates of appreciation during the third year.
There are two main ways a company can re-purchase its shares:
(1) a Tender Offer - where the company offers to buy a fixed number of shares at an above market price. Shareholders can tender their shares to the company, which can scale back acceptances if investors offer more shares than it is seeking.
(2) an On-Market Buy-back - where the company instructs its broker to buy back its shares on the sharemarket over a period of time.
Early research on share re-purchasing - a 1980 study by Larry Dann of 143 tender offers between 1962 and 1976, and another 1980 study by Theo Vermaelen of 131 tender offers from 1962 to 1977 - indicated shares subject to buy-backs did not perform well. Immediately that a tender was announced, stocks rose (by an average of 15%), but did not continue to outperform the market during the following 60 days.
A study by Fortune in 1985 examined 187 buy-backs from 1974 to 1983. From the end of the month of their buy-back through to December 1983 these stocks outperformed the market by 9% per year.
Where the shares were re-purchased in a tender offer the stocks outperformed by 6% per year (following the completion of the buy-back), while stocks subject to on-market purchases outperformed by 10% per year.
Another magazine, Forbes, published a study in 1987 which found that 126 companies re-purchasing their own stock (between 1983 and 1986) outperformed the market by an average of 24% (i.e. about 8% per year).
In 1990 Professor Josef Lakonishok and Theo Vermaelen published another study that examined 258 repurchases made between 1962 and 1986 by all listed US companies. On average these companies offered to buy back 17% of their capital, at a 22% premium to the market and around 85% of shares tendered by investors were accepted.
As observed in the earlier studies, the stock price immediately jumped (by an average of 14%) following the announcement of the buy-back, then only equalled the "market" over the next three months. However, from three months through to 24 months after the buy back was announced the stock outperformed the market by 23% (i.e. about 13% per year).
The study found that the "smaller" the company, the better the performance during this 3-24 months after the re-purchase announcement. Typically, "smaller" companies' shares had been falling sharply for three years prior to the share re-purchase - and their subsequent two year rally dwarfed that of the shares of "larger" companies!
Investment Implications: "Insider" buying and selling by directors is not disclosed in NZ - but this information is available in Australia. Certainly investors should tend to favour Australian shares where several "insiders" have purchased shares during the last year.
Share re-purchases are relatively rare, but can lead to excellent investment returns over the following couple of years - especially among the very "smallest" companies!
In 1967, Robert Levy published a study of Relative Strength
Analysis. Each week for the five years from 1960 to 1965 he
ranked 200 NYSE shares by the percentage amount that the current
price was above or below its average price for the previous 26
weeks (i.e. he compared the current share price to its "26
week moving average").
A strategy of buying shares in the top 10% and selling when they fell out of the top 80% outperformed the market by about 9% per year.
Another strategy - buying shares in the top 5% and selling when they fell from the top 70% - outperformed by about 15% per year.
Norman Fosback's 1976 book, "Stock Market Logic", included the results of his research into Relative Strength on over 750 AMEX listed stocks over an eight year period from 1963 to 1971. Fosback calculated a "strength rating" - being the percentage change in a stock's 30-week moving average over the previous thirteen weeks (i.e. the 13-week change in the 30-week moving average).
Ranked by their strength rating, and divided into five equal groups, the "strongest" shares outperformed by 5.1% per year, while the "weakest" shares underperformed by 5.6% per year.
O'Shaughnessy's recent work (which measured relative strength simply as the percentage change in a stock's price over the previous 12 months) confirms the predictive value of Relative Strength.
The "strongest" shares (from the whole market) outperformed by only 1.5% per year, but the "weakest" shares underperformed by 10.7% per year.
Selecting from only "larger" companies, the "strongest" shares outperformed by 5.1% per year while the "weakest" shares under-performed by 2.3% per year.
O'Shaughnessy also found that relative strength significantly improved results in multi-factor selection methods.
Investment Implications: There is a tendency for share prices to move in "trends". So investors should generally buy into (and, more importantly, hold onto) shares that are rising in price. Similarly investors should generally avoid buying into companies whose share prices are falling rapidly.
Next month, in the final of this series on Share Selection
Methods, we shall use the information presented above to
subjectively formulate a comprehensive Share Selection Criteria.
Although the selection and weighting of the indicators will require subjective judgement, all of the share selection methods we have discussed involve objective numbers that can be calculated using a formula (e.g. the P/S ratio, or a Relative Strength Rating) or by direct observation (e.g. counting up the number of "insiders" buying or selling over the last year).
Securities Research Company maintains computerised databases of all listed NZ companies and all listed Australian companies, so by writing a program to match the comprehensive share selection criteria we shall be able to produce a selection of possible "buy" candidates and a selection of shares that possibly should be sold.