The idea of achieving merely "average" returns isn't particularly palatable to many people, especially when markets aren't performing particularly well. The fact that the "average" investor actually does worse than the broad market indexes and thus achieving a market rate of return is considered to be pretty good doesn't make the returns any easier to accept.

Realistic or not, people want higher returns and as a result they resort to following a wide variety of methods that could only be described as investment quackery. There are many charlatans willing to sell the dream of easy outperformance, but the primary beneficiaries of this advice are the people selling it, not the people following it.

Investment is a mixture of both art and science, but there is a logic to it. The remainder of this article describes methods and asset classes that have been shown to offer a high chance of superior returns.

Value stocks vs. growth (or "glamour") stocks

For nearly one hundred years, there have been a few investors who have quite consistently outperformed the general market averages. Some of them are very famous, even if you aren't particularly interested in investment you still may have heard of Warren Buffett, the "Sage of Omaha", currently the world's richest man. Warren Buffett is a unique billionaire, he is one of the few self made billionaires who got that wealthy as a full time stock market investor.

Warren Buffett calls himself a "value investor". A value investor is someone who looks for stocks which are trading at prices lower than their intrinsic value on the stock market. There are a number of ways you can go about being a value investor, but generally speaking value investors tend to buy shares trading at either absolutely low or relatively low prices compared to their earnings, sales, assets, cash flow etc.

Warren Buffett was a student of Professor Benjamin Graham, author of The Intelligent Investor and co-author (with David Dodd) of Security Analysis. These two books are so highly regarded they are often referred to as "the Bibles of value investing" (the former being a more general book suitable for small investors, the latter being a much more technical treatise on how to pick apart company accounts.)

Benjamin Graham was more than just an academic, he was also a highly successful money manager. His investment company, the Graham-Newman Corporation, achieved returns of more than 20%pa over a 20 year period, substantially outpacing the general market.

To commemorate the 50th anniversary of Security Analysis, on 17 May 1984 Buffett gave a seminar at the Columbia Business School (where Graham taught) called "The Superinvestors of Graham and Doddsville" where he recounted the exploits of a number of value investors and argued that the continuing success of value investors cannot be explained by luck alone, he argued that there was "something in the water" of "Graham and Doddsville" that resulted in an unusual concentration of success among practitioners. This speech in edited form appears as an appendix in recent editions of The Intelligent Investor.

You can read a .html version of Buffett's speech here, a scanned .pdf copy of the original article printed in Columbia Business School's Hermes magazine is here . I suggest you read this article now, because it sets the stage for the commentary that follows.

Buffett's list of Superinvestors is not a complete list of all the top money managers, there have been a number of value investors who enjoyed very long term success over careers lasting several decades, other top value managers include Sir John Marks Templeton and John Neff .

All of Buffett's "Superinvestors" outperformed with quite some consistency over careers lasting decades. Note the Sequoia fund for example, (run by Bill Ruane), it closed to new investors in 1982 because it was flooded with money after more than a decade of fantastic performance, but they've continued to outperform the market by a handy margin since the closure. Sequoia wasn't chosen with the benefit of hindsight from a large list of managers, Ruane was the one and only manager that Warren Buffett recommended to the investors in his first investment partnership when Buffett chose to wind that partnership up in 1969.

Another one of the "Superinvestors" is still in business, Tweedy Brown have a number of articles summing their collected insights into how to beat the market. They incorporate all of this into their own investment style.

Prior to the 1990s, academia basically ignored or tried to explain away Warren Buffett and his value investing peers. Benjamin Graham referred to some early studies into the outperformance of value stocks in his books and a smattering of papers appeared over the years all claiming that Graham and Buffett were right, but academia were entranced by the Efficient Markets Hypothesis (EMH). This theory makes the claim that it is impossible to outperform the market consistently without simply taking on more risk (like by gearing the portfolio) or by sheer luck. Since most of the "Superinvestors of Graham and Doddsville" didn't believe in using gearing, academics argued that they must just be the product of pure luck. Warren Buffett, they said, was an extraordinarily lucky man who happened to be successful at guessing the outcome of a coin flip - sooner or later his luck would run out.

In 1992, academia finally began to take value investing seriously when two of their most respected members, Professors Eugene Fama and Kenneth French wrote a paper (Fama, Eugene F., and Kenneth R. French, 1992, The cross-section of expected stock returns , Journal of Finance 47,427-465.), in which they performed a statistical analysis of the US stock market from 1963 to 1990. In this study they found that, contrary to expectations, as a matter of fact there was a "value premium", as well as a premium for purchasing small companies. They proposed a model of stock returns that they refer to as the "Three Factor Model ". The Three Factor Model explains portfolio returns as a combination of a "market" factor, i.e. how much of the portfolio is invested in stocks versus bonds, a "size" factor, i.e. whether you are buying large or small companies, and a "value" factor, i.e. whether you are buying stocks that are cheap or expensive with respect to quantitative measures of value, in Fama and French's case the relationship of price to book value.

Diagram by Dimensional Fund Advisors, used with permission.

Ken French's web site has a data library section where historical data files going back to 1926 are available for free download.

There is a fund manager that closely aligns itself with the work of Fama and French, Dimensional Fund Advisors. It has a number of excellent articles on financial theory, including a very good historical article describing the development of financial theory prior to the Three Factor Model Explaining Stock Returns: A Literature Survey and an updated account of the Three factor Model with international as well as American data, The Dimensions of Stock Returns: 2002 Update.

The theory essentially says that if you want better returns you can get them by going for smaller companies and value stocks (note: these are not mutually exclusive, you can buy small value stocks).

Fama and French chose to use as a benchmark of "value" the relationship between a stock's book value and market price. Book value is the value of a company's assets minus its liabilities, it is also known as "shareholders equity". Investors traditionally divide the price by the book value to get a ratio "Price to book ratio", but Fama and French flipped this upside down because some companies have a zero or negative book value, using the inverse measure avoids divide by zero errors and allows one to find "value" by simply identifying the stocks with the highest book to market ratio.

Some people question whether the "book to market ratio" (BtM) is really an appropriate measure of value. Why not compare the price with earnings, cash flow or sales? As a matter of fact Fama and French did try out those other measures, but found the BtM measure was generally superior in that book value is less volatile than earnings. For example, a company can have an enormous price to earnings ratio (generally seen as making it expensive) but if the company had simply scraped along with near breakeven results as part of a temporary glitch then it might actually be considered cheap compared to its normal earnings. There is a good article at the Dimensional Fund Advisors web site called Is There Still Value in the Book-to-Market Ratio?, which addresses this issue.

What was unique about the Fama and French paper was not really what was said, but who said it. These two professors were instrumental in the development of the EMH in the first place, for example it was Eugene Fama that first proposed that stock prices fluctuate randomly and that no system could be made out of price data that would be able to outperform the market. Fama and French have always been among the EMH's most formidable supporters. Fama and French were the last people you would expect to write a paper claiming there were ways to outperform a random selection of stocks.

As one professor commented while discussing Fama and French's "discovery":

"Modern finance today resembles a Meso-American religion, one in which the high priest not only sacrifices the followers - but even the church itself. The field has been so indoctrinated and dogmatised that only those who promoted the leading model from the start are allowed to destroy it."

As soon as the paper was published, it came immediately under heavy attack from all sides. Fama and French were accused of "data mining", using tainted data, using less than rigorous statistical methodologies and far worse. A series of papers were written making these accusations, but they were comprehensively refuted over the next few years as other researchers replicated the work.

The Fama/French work could have been rebutted if it were found that in other markets, or on other time frames, the model didn't work. In what must have been a dark time for EMH traditionalists, one paper after another was written confirming the existence of a "value premium" on different time frames and in different markets. If the value premium were just a lucky twist of fate on the US market in the 1963 - 1990 period then it should not have been repeated elsewhere.

When Fama and French's work was extended back to 1926 the value premium was observed in the earlier period as well, and then a number of papers emerged documenting the same thing in various European and Asian markets.

Three papers that document international value premiums are here, here and here.

The graph below (which I prepared using Fama and French data provided by Dimensional Fund Advisors in the international edition of their Returns software) charts the performance of Fama and French's large cap data series (on the US market) commencing July 1926 until March 2004 over rolling five year periods.

This next chart shows the same information for Fama and French's small company indexes.

If you extend the holding time to 10 year periods, the "value premium" becomes even more consistent. Here are 10 year rolling periods for large companies...

... and small companies.

Over the entire period, the annualised returns of each of the indexes appears to neatly follow the Fama/French Three Factor model, tilts toward both value and small companies tend to produce higher returns.

Similar results have been established for a variety of international markets, including Australia. In fact, Australia appears to have had one of the largest value premiums that I've seen anywhere. From January 1980 to April 2004, a value index constructed using the Fama/French methodology has outperformed the broad benchmark S&P/ASX500 All Ordinaries Accumulation index by more than 7%pa, returning 19.61%pa against the All Ordinaries 12.60%pa. However, Australian small companies have not performed particularly well, despite achieving very high returns in 2003 they still underperformed over the whole period, returning 11.70%pa. As with the US data, my source of this returns data was Dimensional's Returns software (Australian edition).

This next chart shows the annualised return of a value index and a small company index vs. the All Ordinaries over rolling 5 year periods.

A paper from the Brandes Institute on international value premiums (The Value Premium in Non-U.S. Markets, October 2003) did provide some measurements of Australian value vs. growth investing for large and small stocks. Although small companies have underperformed large companies, the difference in return between small cap growth and small cap value in Australia is enormous. Obviously one should be extremely cautious if considering buying a high priced small cap stock in Australia.

Chart from The Brandes Institute, The Value Premium in Non-U.S. Markets, October 2003

Chart from The Brandes Institute, The Value Premium in Non-U.S. Markets, October 2003

Source data: The Brandes Institute, The Value Premium in Non-U.S. Markets, October 2003

The EMH counterattack - are value stocks more risky?

Early critics of the work on the value premium at first sought to dispute the data claiming the calculations were flawed, the data biased or it was mere statistical accident. As data accumulated demonstrating how widespread and consistent value premiums have been these criticisms were for the most part set aside. Efficient market hypothesis believers, who cling to the idea that one can only outperform by taking greater risk, argue that the value and small cap premiums are the result of higher risks and the three factors of the Fama/French model are in fact three risk factors.

Fama and French themselves are among this camp, they propose that value stocks are more risky than growth stocks, hypothesising that companies which trade at low prices must be doing so because the market has discounted their prices because of high risks. Several theories have been put forward to describe what this risk must be, but so far there is little convincing empirical evidence to prove the existence of this "value risk".

If the reason why value stocks outperform growth stocks is because they are more risky, it is very difficult to see this in terms of the volatility of the indexes themselves. Value indexes are notoriously less volatile than growth indexes, and tend to fall less during bad markets. This makes some sense because, as Benjamin Graham pointed out so many years ago in The Intelligent Investor, value stocks trade at low multiples because the market does not expect much in the future for them whereas growth stocks trade at high multiples because the market expects a lot.

Given the way that stocks are valued as the present value of all future cash flows, a much greater portion of the price of "growth" stocks represents expectations of the future, which are likely to be rated very differently as the market's appetite for risk waxes and wanes. To the extent that value stocks trade on what they are worth right now in terms of assets and dividends, whereas growth stocks trade on what people hope they will be worth if they achieve extraordinary feats, it makes sense that growth stocks should be (and are) more volatile.

My favourite alternative measure of past risk is to look at the "drawdown" of an index. The drawdown is a good way to chart losses, it charts how far an index (or any asset) has fallen relative to the highest previously achieved high. If value stocks are more risky than growth stocks, the higher risk doesn't appear to show up in drawdown charts.

Value (both large and small) did stumble on their way to recovering from the Great Depression, almost "double dipping", but aside from only brief episodes ever since value has offered superior protection during bear markets. The Great Depression was a unique bear market where even the highest quality stocks got knocked down to remarkably cheap prices. Given a choice between buying high quality companies at extraordinarily inexpensive prices or lower quality companies at slightly lower prices, most investors bought quality. Other bear markets since then have seen value stocks outperform.

At worst, you could say that value was about as risky as growth, you can't conclude by looking at the drawdown figures that value was more risky than growth. There is certainly no extra risk in these charts obvious enough to account for a 2%pa level of outperformance in large companies and a 5%pa level of outperformance in small companies.

Especially in American small caps, growth tended to fall further and take longer to recover from losses than value. It is clearer in small caps that small cap growth has been more risky than small cap value, despite the significantly better performance of small cap value.

Unfortunately, the small size of the Australian market does not lend itself to splitting up value into large and small companies (well they could, as the Brandes Institute did, but the indexes wouldn't have very many stocks in them and passive managers like Dimensional prefer diversification), so Dimensional produced a single "value index" which incorporates both large and small value companies, they also produced a neutral small company index but didn't split that into value or growth.

Clearly small companies have had more severe drawdowns than the other indexes but when you compare the Australian value index with the All Ordinaries it is not obvious that the value index, at least over this 28 year period, was any more risky than large companies.

Value and the Great Depression

The inferior performance of value indexes during the 1930s is often cited as evidence that value stocks really are more risky. The theory, which appears to be based on little more than just how value compared to growth through the 1930s, says that value companies must be exposed to economic risk to a greater extent.

It is hypothesised that deflation, which was severe during the depression, is a risk factor that hurts value more than growth. EMH writers keen to dispell the "myth" of a risk free value premium warn that if America ever goes through another period of deflation, value investors will suffer greatly.

I have a number of problems with that theory. As shown in the next two charts, during the initial crash and subsequent bear market from September 1929 to June 1932 value stocks didn't actually fall any more than growth stocks, they just took a bit longer to recover from the fall. Also, the period of underperformance was so brief I'm inclined to discount it as random noise rather than read anything into it. When value did recover it did so convincingly and substantially outpaced growth during the decades that followed. The following are cumulative return charts for the US Fama/French indexes from January 1929 to December 1946.

Another objection I will raise is that in another more recent major deflationary bear market, the one in Japan that lasted through the whole 1990s, value outperformed growth. Theories that claim value must be more risky than growth during deflationary times should account for why Japan's value premium was so strong during the 1990s. The value premium in Japan was so strong in fact that value actually delivered positive absolute returns during what most people would recall as a severe bear market.

This chart shows the cumulative percentage returns of the Fama/French high BtM value and low BtM growth index in Japan from 1975 to 2003. Note the logarithmic scale which understates the amount of outperformance.

This is a chart of rolling 5 year returns for a Japanese high BtM (value) and low BtM (growth) Fama/French index.

This is a drawdown chart based on the same data.

And this last chart shows the return of the Japanese Fama/French value and growth indexes if bought at any time in the last 28 years and held until December 2003. As you can see, despite the "bear market" in Japanese equities from 1989 until 2003, the Fama/French value index actually made positive returns in Japan no matter when you begin your holding period. The growth indexes are still 70% below their 1989 highs but value investors have made an annualised return of about 4.88%pa since January 1989, or 104.4% cumulative. (Note: returns on this chart from January 2003 to December 2003 are actual, returns from periods prior to January 2003 have been annualised.) The value premium was not only present in Japan's deflationary market, it was very strong and consistent.

My last objection is to the assumption that the efficient market hypothesis can be extended back that far. Graham and Dodd basically invented fundamental analysis with their landmark 1934 book Security Analysis, prior to Security Analysis very few people had any notion of intrinsic value, what drove prices was pure speculation, insider trading, deliberate market manipulation, chart reading and rumours. In those days companies didn't even disclose basic information about their profitability, assets and sales. Traders had to act on hunches and rumours or even industrial espionage. Most of these are still major influences today, but at least these days some investors use a value approach, financial data is quite well disclosed and the idea of buying stocks based on fundamental analysis is no longer novel. The notion that the market was "efficient" in the sense that stocks were priced appropriately under those circumstances would have amused Benjamin Graham enormously. If the market probably wasn't efficient then basing an EMH risk hypothesis on the behaviour of low priced stocks during that time is a bit of a stretch.

One major common factor that binds the 1930s to today's markets is that human nature is still basically the same - and, as I argue below, it is human nature which many people believe causes the value premium. The claim that the experience of the Great Depression proves the riskyness of value stocks does not seem to be soundly based.

EMH counterattack number two: risks are fundamental

In response, the defenders of the EMH have proposed that some forms of risk aren't visible in prices, they argued that value is riskier than growth because value companies are "financially distressed".

Again, there are problems with that theory. A recent paper (Griffin, J. M. and M. L. Lemmon. "Book-to-Market Equity, Distress Risk, and Stock Returns." The Journal of Finance, Vol. 57 No. 5 (2002): 2317-2336.) measured distress risk in value and growth companies, using a measure called "Ohlson's O-score", which is a measure of the likelihood of bankruptcy. They found that among the firms with the highest O-scores were many "value companies", but an even greater number of "growth companies". They found that the growth companies with the highest O-scores had extremely low returns but this distress risk was already priced into "value companies" and hence filtering value by their O-score didn't add any meaningful information. According to the authors, "These large return differentials cannot be explained by risk as captured by the Fama and French three-factor model, nor differences in economic fundamentals, such as profitability or the likelihood of delisting. In contrast, predictions of the overreaction hypothesis are borne out. Distressed firms exhibit the largest return reversals around earnings announcements, and the book-to-market return premium is largest in small firms with low analyst coverage."

Another paper, Dichev, Ilia D., "Is the Risk of Bankruptcy a Systematic Risk?" Journal of Finance, Vol. LIII, No. 3, University of Michigan Business School, (Jun-1998), pp. 1131-1147 has this to say:

Several studies suggest that a firm distress risk factor could be behind the size and the book-to-market effects. A natural proxy for firm distress is bankruptcy risk. If bankruptcy risk is systematic, one would expect a positive association between bankruptcy risk and subsequent realized returns. However, the results demonstrate that bankruptcy risk is not rewarded by higher returns. Thus, a distress factor is unlikely to account for the size and the book-to-market effects. Surprisingly, firms with high bankruptcy risk earn lower than average returns since 1980. Additional results suggest that a risk-based explanation cannot fully explain the anomalous post-1980 evidence.

William Bernstein wrote an article in his Efficient Frontier.com website summarising several papers on the search for a risk explanation of the value premium called "Tastes, Distress, and Jocks" in which he discusses several papers which attack the risk explanation:

In a recent working paper John Campbell and his colleagues looked at metrics of market distress and their predictive value, both in terms of subsequent bankruptcy and returns. Without going into all the gory details, the authors identified several new balance-sheet ratios suggestive of company distress that did a dandy job of predicting future bankruptcy—much better than the traditional techniques. The Fama-French risk hypothesis predicts that distressed companies identified by these techniques should have higher returns than the market. Alas, no: The most distressed companies had returns that were much lower than those of the least distressed companies, with multifactor alpha spreads on the order of 20% per year. About the only way an efficient-market enthusiast can wiggle his way out of this one is to posit dimensions of risk beyond company failure—a tall order (or else yell "data mining!" at the top of his lungs).

Another paper from Australia is The relation between distress-risk, B/M and return: Is it consistent with rational pricing?.

Fama and French are also writing a paper, due in early 2008, where they examine how corporate financing decisions correlate to the company's Book value to Market ratio. The paper is built on the assumption that where managers consider their companies to be overvalued they will seek to profit from this by selling their expensive equity and raising debt at favourable terms while they can, they will tend to seek finance by selling equity rather than debt, and longer term debt rather than shorter term debt, and pay dividends rather than buy back stock.

Conversely, where managers consider their stock to be undervalued they would tend to raise finance by borrowing rather than selling equity, and when they do borrow they should borrow at short terms rather than long, and would use excess capital to buy back their undervalued stock rather than pay dividends.

Noting that there are several confounding factors which would work to hide a positive result, such as executives being more likely to pay themselves cash bonuses rather than equity when they consider the equity to be overvalued, and for undervalued companies to use high dividends as a means of adding some support to the stock price, Fama and French still found evidence that corporate financing decisions indicate that managers of "value" companies do consider their stocks to be undervalued and managers of "growth" companies tend to consider their stocks overvalued. That this effect is visible and statistically speaking quite strong, despite the confounding factors indicates that there is a high degree of managers disagreeing with the valuations placed on their companies.

In recent years Fama and French have become increasingly receptive to the hypothesis that the value premium is at least in part driven by mispricing rather than pure risk. At this point the main question they are looking into is trying to figure out what proportion of the value premium is risk, and what is return.

There have been attempts to improve on basic value filters by applying traditional fundamental analysis techniques to remove risky securities from the indexes. Devising strategies which produce a higher return than vanilla value and growth indexing by removing fundamentally riskier holdings provides further evidence that the value premium is not driven by risky companies, but by undervalued companies.

Piotroski of the University of Chicago came out with an interesting paper in 2000 where he showed that some common fundamental analysis strategies which would be familiar to Graham and Dodd devotees can screen out losers from value indexes, resulting in a massive 7%pa outperformance of a value index.

This paper has been frequently cited and it was commented on by Guay and in a followup paper by Mohanram that looked at growth indexes and also found that fundamental analysis could improve on the index return, weeding out losing stocks. Josef Lakonishok is another leading academic who has done a lot of research on the value premium who strongly disagrees with Fama and French's explanation that value stocks are more risky. Lakonishok, together with Louis K. C. Chan wrote a very good review paper on value investing in July 2002, Value and Growth Investing: A Review and Update. Lakonishok takes the view that value outperforms growth because the market simply overestimates the future profit of growth stocks, extrapolating high levels of historic growth too far into the future and underestimates the future profitability of value stocks. Lakonishok and Chan are as scathing of Fama and French's value risk theory as one can be while still adhering to the polite language of an academic journal:

"Fama and French (1996) argue that stocks with high ratios of book equity to market value are more prone to financial distress and hence riskier. They employ a version of the Merton (1973) multi-factor asset pricing model to account for value stocks' higher risk exposures to a financial distress factor, and hence their higher returns. This argument, however, stretches credulity. On the basis of the risk argument, it would follow that Internet stocks which had virtually no book value but stellar market values were much less risky than traditional utility stocks which typically have high book values of equity relative to market. It is also noteworthy that the idea that value stocks have higher risk surfaced only after their higher returns became apparent. Data snooping is considered to be a sin, and coming up with ad hoc risk measures to explain returns should be regarded as no less of a sin."

To save the EMH value risk story, EMH apologists are faced with a daunting task, how do you explain to jeering crowds of value investors that the Dot Com stocks, which as we all know nearly all went broke, were less risky than the much more profitable "old economy" stocks that value investors tended to buy during the tech boom?

The most glaring problem with the assertion that growth stocks are less risky than value stocks is the recurring phenomenon of price bubbles. Investors today have painful memories of the Nasdaq bubble of 1998 to 2000, where a number of loss making companies involved in highly risky and unproven internet businesses were briefly given values far exceeding those of many large and well established and highly profitable "old economy" companies.

The definition of a "growth" stock used by academics studying the "value premium" is simply the price to some accounting number such as earnings or sales or most commonly book value. The size of a company is measured by its market capitalisation, which is simply the value of all the company's shares (number of shares times share price). Therefore, a "large growth company" is a company with a very high price in relation to assets and a large market capitalisation.

At their peaks, many of the "Dot Com" stocks hit market capitalisations and valuations that put them high up in the range of "large cap growth" stocks. If the value risk theory holds this means they are very low risk. In fact, the higher the prices got the larger their market capitalisations and the more "growthy" they became. If the EMH is valid, in early 2000 the lowest risk stocks on the planet were Dot Com companies which had only ever lost money. These were far less risky than the "old economy" stocks.

In April 2000 the Nasdaq started falling, Dot Com companies found it difficult to obtain financing to cover their enormous losses and the majority of them went broke. Even among the technology companies that were actually profitable such as Microsoft and Cisco losses were terrible.

On the other hand, many "old economy" companies, now that investors were able to pry their eyes away from the technology sector, actually performed very strongly and not only continued to make very good accounting profits but also did very well on the stock market.

Warren Buffett, who had attracted enormous criticism during the tech boom for his quaint ideas about investing only in profitable established companies at attractive prices, prospered enormously during the "bear market", the price of Berkshire Hathaway doubled at the same time that the Nasdaq experienced falls not dissimilar to those experienced by investors in the 1930s during the Great Depression.

It is all very well to claim that the Dot Com boom was somehow a one off, or that the outcome couldn't be predicted, or that perhaps it could but was "rational", as many EMH supporters claim, but one of the most arkward facts that confront the EMH is that booms like the Dot Com boom happen quite frequently. Personal computers had a similar stock market boom and bust during the 1980s, the 1960s had a particularly strong boom in electonics which busted in the early 70s and the Dot Com boom also has eerie similarities with the 1920s boom in radio stocks (which were at that time no less revolutionary and miraculous than the internet).

In fact, although it isn't always technology (sometimes it is gold or oil stocks or real estate), major booms and busts in one glamour sector or another of the stock market have on average occurred more than once a decade through the whole 20th century, and this has been going on for centuries.

It is always hard to know when these booms will end, and always very tempting to get in on them when they are in progress, but markets have a very long track record of going way overboard in one sector or another and the mess is always the same following it. This recurring phenomenon has been known about for centuries. One of the most readable books on this subject is A Short History of Financial Euphoria by respected Harvard economist, John Kenneth Galbraith. Galbraith talks about numerous episodes in financial euphoria from the 18th century to the 1980s, pointing out many common elements and proving the old adage that "the one thing we learn from history is that we do not learn from history".

Galbraith wrote this book in 1990, but it was by no means the first book written on bubbles and euphoria nor, apart from it being one of the best written, did it really come up with any startling new insights. The subject of bubbles was already old hat by the time Benjamin Graham wrote The Intelligent Investor.

One of the reasons why value investors claim their method is less risky is that by defininition their method avoids the most expensive sectors of the market and thus keeps them out of bubbles.

So what causes the value premium?

There have been several papers which have documented a puzzling fact: often you get much better performance by investing in "bad" companies than in "good" companies.

In 1982, one of the best selling business books was In Search of Excellence: Lessons from America's Best Run Corporations by Tom Peters and Bob Waterman. This book extolled the virtues of a number of very well run companies in the US, and set out a number of criteria for quantifying the quality of management.

Michelle Clayman wrote a paper five years later which compared the performance of "excellent" companies with a set of "unexcellent" companies which rated very poorly by the same criteria in the years since the book was published. (Clayman, M. (1987), In search of excellence: The investor's viewpoint, Financial Analysts Journal, May-June, 54-63.). Her findings were that the "excellent companies" did not perform significantly differently from the S&P500 index, whereas the portfolio of "unexcellent" companies outperformed (sic) the market by 12%pa.

Another study (Kolodny, R., M. Laurence and A. Ghosh (1989), In search of excellence… for whom?, Journal of Portfolio Management, 15 (3), 56-60.), using more detailed analysis found no significant difference in performance between Peters and Waterman excellent firms compared with either the market index or an appropriate control sample.

A new study, not yet published, has been presented at several conferences by Vineet Agarwal of the Cranfield School of Management in the UK (also authored by Mike Brown of the Nottingham Business School and Professor Richard Taffler of the Cranfield School of Management). This paper, titled "Are Well Managed Companies Good Investments" performs a similar analysis based on the UK publication Management Today's listings of "Britain's Most Admired Companies". Their data seems to support a conclusion that the stocks of Britain's most admired companies do not outperform perform Britain's least admired companies. Furthermore, they find that the rankings of admired companies seem to be heavily influenced by past growth and stock market performance.

Another study, which wasn't particularly controversial because it basically repeats a claim often made by some types of EMH theorists is that the earnings growth rates of companies lack persistence and predictability, though there is some persistence in sales growth. They also find that value multiples such as book to market add little to the predictability of earnings (which means that "value" companies grew at about the same rate as "growth" companies after portfolio formation). This means that there is little justification in awarding high prices to companies with very strong track records - the growth rate on which the stock's high valuation depends is not likely to persist. (See Chan, L.K.C., J. Karceski, and J. Lakonishok. "The Level and Persistence of Growth Rates." Journal of Finance, Vol. 58 No. 2 (April 2003): 643-684.)

To sum up Lakonishok's major finding:

While some firms have grown at high rates historically, they are relatively rare instances. There is no persistence in long-term earnings growth beyond chance, and low predictability even with a wide variety of predictor variables. Specifically, IBES growth forecasts are overly optimistic and add little predictive power.

Cragg and Malkiel (the latter being the author of A Random Walk Down Wall Street) did an early analysis of long-term estimates, published in the Journal of Finance (23, March 1968, "The Consensus and Accuracy of Some Predictions of the Growth of Corporate Earnings" pp 67-84), looking at the projections made by groups of analysts at five respected firms, covering 185 stocks. The researchers found that most analysts estimates were based on linear extrapolation of current trends with low correlations between actual and predicted earnings.

They found that analysts would have substantially improved their accuracy if instead of extrapolating past growth rates they had simply inserted the long term company average growth rate of 4% annually.

Another study, by Oxford professor I.M.D. Little ("Higgledy Piggledy Growth", Bulletin of the Oxford University Institute of Economics and Statistics, November, 1962) found that corporate earnings in fact seemed to follow a random walk, with little correlation between past and future rates. Recent trends provided no insights useful for forecasts.

As Benjamin Graham wrote in Security Analysis:

The truth of our corporate venture is quite otherwise [than investors think]. Extremely few companies have been able to show a high rate of uninterrupted growth for long periods of time. Remarkably few also of the large companies suffer ultimate extinction. For most, this history is one of vicissitudes, of ups and downs, with changes in their relative standing.

So it is dangerous to forecast that high rates of return will last for ever, just as it is foolish to assume the worst will happen all the time. In practice analysts have only a very limited ability to anticipate future profits, yet their forecasts carry such weight that they are able to drive a wedge through the market, undervaluing some stocks and overvaluing others. If high growth rates are very difficult to sustain and even harder to predict then it would be irrational to pay a significantly higher price multiple for a stock based on anticipated growth. In reality, investors do pay very high multiples for expected growth and this is, to many researchers, the simple explanation for why there exists a value premium.

What do these studies showing that "excellent" companies do no better than "unexcellent" companies, and that high growth rates are no more likely to persist than low growth rates tell us?

I'll let another great value investor, Sir John Marks Templeton explain in his own words:

"It is crucial to understand, and very few people do, that attaining superior investment performance has nothing at all in common with succeeding in 99% of other occupations. If you were building bridges and a dozen consulting engineers experienced in bridge building all gave you the same advice, you'd be stupid not to build your bridge their way. In all probability, if the experts all agree, their way is the right way to do it. You'd build a better bridge at lower cost if you followed their advice. But the very nature of the investment-selection process turns that scenario topsy-turvy. Let's assume that every securities analyst you see says, 'that's the stock to buy!' You might think that if all the experts are saying "buy", you should. But you couldn't be more wrong. To begin with, if they all want it, they'll probably all buy it and the price will build up enormously, probably to unrealistic levels. By the same token, if all the experts say, 'it's not the stock to buy,', they won't buy it and the price will go down. It's then, if your research and common sense tell you the stock does have potential, that you might pick up a bargain.

That's the very nature of the operation. It's quite simple; if everybody else is buying, you ought to be thinking of selling. But that type of thinking is so peculiar to this field that hardly anybody realises how valid it is. They say: 'I know you're supposed to look where other people aren't looking,' but very few actually understand what that means."

(From: Global Investing: The Templeton Way by Norman Berryessa and Eric Kirzner (Irwin Professional Publishing, 1993))

Yes, that's right, the advice you have been given for years and years that you should only invest in well managed, solid "blue chip" companies was, if the person giving it didn't add the words "as long as they cost no more than bad companies", poor advice. Every article you will ever read in investment magazines and books, every report written by fund managers, every well meaning tip from friends and relatives has been basically wrong.

If the market is reasonably efficient, the quality of a company will already be built into the price. There is no reason to believe that buying better companies will give superior results unless you know that the market is unaware of the company's quality. I lost count in the late 1990s of the number of stock broker reports that came across my desk recommending widely recognised stocks at stratospheric prices because they were "premium quality blue chips" with a strong record. Most of those premium priced "premium quality" stocks suffered "premium size" losses when the bear market hit because the market had already put all of this "quality", and then some, into the price.

But the evidence seems to point not just to quality being built into prices, many researchers believe that quality is overbuilt into prices. The value premium is, according to this theory, the result of people overpaying for quality companies and overlooking lower quality companies, regardless of how attractive they may be on a valuation basis. This is known in the field of behavioural finance as the "representativeness heuristic".

A "heuristic" is a mental shortcut or stereotype we use to simplify calculations and speed up decision making. Some of these shortcuts are sensible and do help us, for example although not all spiders are dangerous it is safe for us to assume that all are and try to avoid contact with them, this saves us from the occasional bite by an actual dangerous spider. It is faster for our brain to run on heuristics and generalisations than to have to catalogue each variation on a theme. Some heuristics are not helpful though, including most investment heuristics because as Templeton explained above investment is by its nature counterintuitive. The "it is a good company so it must be a good investment" heuristic is dangerous and frequently wrong. So also is the "it is a bad company, so it must be a bad investment" heuristic.

This misleading heuristic is so powerful, and so ingrained, that investors can be fooled by it again and again and again, repeatedly losing money by paying through the nose for companies on the basis of a favourable review in the newspaper or dumping stocks after a bad review, only to see the stock they just bought fall in price or the stock they just dumped rise. There are some types of investors, called "distressed equity" managers, who make spectacular returns by investing in companies on the verge of, or actually in, bankruptcy, they make extremely high returns by buying "bad companies".

Another author who argues that the value premium is the result of systematic biases in forecasts is Robert A. Haugen, author of The New Finance : The Case Against Efficient Markets. Haugen's research has found that while "growth" companies do still grow their earnings at a faster rate than "value" companies, at least for a few years after being identified as a "growth stock", the premium paid for growth stocks is almost always too high because the period of higher growth doesn't last long enough to justify the higher price.

One of the most interesting books written on value investing is Contrarian Investment Strategies: The Next Generation by Forbes columnist and fund manager David Dreman. I summarised this book in an article in the shares section of my FAQ.

Dreman was one of the early writers on the value premium (apart from Ben Graham of course), and was often singled out by EMH academics as a charlatan for his claims that value stocks outperformed, with many academics seeming to devote enormous amounts of energy to trying to prove his claims wrong. For example, when studies started to show that small companies outperform large companies, the authors quickly rushed to the conclusion that this disproved Dreman's claims about value, there was a small company premium and therefore Dreman's theories about value outperforming were wrong. (!)

For example, Fortune magazine, then a bastion of EMH thought, had this to say after a 1980 study which had discovered that small companies outperform:

[T]he small-stock phenomenon has indirectly refuted the most serious challenge yet to the efficient-market theory. A number of researchers have demonstrated that portfolios of stocks with low price-earnings ratios have regularly outperformed the market averages. That finding, trumpeted by David Dreman in his book Contrarian Investment Strategy, is wholly inconsistent with an efficient market. It turns out, however, that low P/E stocks appear to offer superior returns only because small stocks have lower P/Es, on average, than large ones.

In time, Dreman's views have come to be vindicated by researchers but his published papers remain obscure (to the academic community) and are rarely referenced. Dreman documented a very interesting phenomenon with value and growth stocks when he, along with his colleague Michael Berry, undertook a very comprehensive investigation into analyst forecasts, their accuracy and their effect on prices. Other studies have been done on this subject of course, but Dreman's work was interesting because he divided up the stock universe into low priced "value" and high priced "growth" stocks to see if analyst expectations were systematically different.

Analyst errors and "earnings surprises"

First of all, Dreman found that the error rate in analyst forecasts was unacceptably high in all cases. Stocks can rise or fall strongly if a company beats or underperforms market expectations by only a few percent, but Dreman found the errors were much higher than this.

The Dreman/Berry study was one of the largest studies of broker's quarterly earnings forecasts ever done. (Dreman, D.N., and M.A. Berry. "Analyst Forecasting Errors and Their Implications for Security Analysts." Financial Analysts Journal, May/June 1995a, pp. 30-41.)

This study examined brokerage analysts' quarterly forecasts of earnings as compared to earnings actually reported between 1973 and 1991, which has subsequently been extended to 1996. Estimates for the quarter were usually made in the previous three months, and analysts could revise their estimates up to two weeks before the end of the quarter. In all, 94,251 consensus forecasts were used, and they required at least four separate analysts' estimates before including a stock in the study. Larger companies such as Microsoft or Exxon, might have as many as 30 or 40 estimates. More than 1,500 NYSE, NASDAQ and AMEX companies were included, and on average there were about 1,000 companies in the sample.

Many market professionals believe an error of +-5% is enough to trigger a major price move, so to be of any worth the forecasts must be inside that range. The average error of the 500,000 individual analysts reports in fact was 44% annually. In fact this is an average over the whole time, but for some reason analysts estimates actually got worse over the period, in the 1990s the average error was more than 50%.

They tested for skews in the data. Are the errors inflated by a few very large errors that dominate the sample? No, they had a look and errors were fairly evenly paced out and forecasts were consistently bad. They also tracked errors by industry, maybe the methods work well for some stable industries but not others. Again a skew was not found. Analysts were no better at predicting the earnings of stable blue chip financial companies than they were at predicting the earnings of highly speculative stocks. Every industry they looked at had forecast errors that were far too high, except for tobacco, which was the only industry with a single digit error, being 4%. The next lowest error was 25% in telecommunications and foods. What about small versus large companies? The results were a little better for very large companies, but not much. Errors were still 23% on average, almost 5 times too high to be usable.

If a forecast is to be of any use at all it must be within 5%, yet the error is creeping up to more than ten times that amount, one might ask what use at all is this information? Who would trade on this advice? Obviously following the advice of these analysts is extremely bad for your financial health. But this is exactly how people play the game in the stock markets, experts receive major adulation and billions of dollars are sent after these flaky predictions year after year.

So what proportion of estimates were actually on track? Out of 94,251 estimates Dreman found that 29.4% of them were within plus or minus 5% of actual earnings. Less than half (46.8%) in fact were within plus or minus 10% and only 58% of consensus forecasts were even within 15%, a tolerance level that most Wall Streeters would agree is far too high. This creates a serious problem, because companies are not sold on one year forecasts, but on consensus estimates of profits many years into the future. The forecasters were no better at predicting long term earnings than short term, meaning that if taken as a whole the investor who invests consistently in broker recommended stocks has a cumulative probability of beating the market of next to nothing.

Dreman tested results over booms and recessions, and found no large difference in accuracy for different economic conditions. In all cases analysts were off.

Which direction were the analysts errors? Another study is cited where Jennifer Francis and Donna Philbrick examined analyst estimates from the Value Line Investment Survey, 918 stocks for the 1987 - 1989 period and found that analysts were on average too optimistic, overestimating by approximately 9%. In a report to subscribers, IBES, the largest earnings forecasting service which monitors earnings on over 7,000 companies found that the average revision to forecasts for companies in the S&P500 is 12.9% from the beginning to the end of the year in which the forecast is made. Analysts revise their estimates 6.3% in the first half and 19.5% in the second half of the year. Despite the changes, however, analysts recommendations seem to still be on average far too optimistic, they revise their optimistic estimates at the start of the year and triple that revision, usually downwardly in the second half, yet after all this they are usually still too optimistic.

Lakonishok also found in his study on the level and persistence of growth rates that forecasts tended to be too optimistic on the whole.

The effect of forecasts being too optimistic is that stocks tend to fall when the final results are announced. If you want to use analyst forecasts, you should make a habit of adjusting them downwards.

But Dreman's most interesting finding was that when you check forecast errors in groups of "value" and "growth" stocks, the errors tend to go in the opposite direction. Analyst forecasts are far too optimistic for growth stocks but they are actually on the whole too pessimistic for value stocks.

Even more interesting was what happened when results were announced. Dreman found that earnings releases resulted in significant falls for growth stocks but significant rises for value stocks. When a growth stock outperformed analyst expectations (which occasionally they did), the stock didn't respond to the good news. The reason for that was because the market had such inflated expectations to begin with that the market seemed to expect the company to beat consensus forecasts. On the other hand, if a growth stock delivered less than market expectations the price would react violently with a strong selloff.

The opposite actually happened with "value" stocks. When a stock widely regarded as a "dog" delivered a result below forecasts the market tended to take this in its stride. The market had such low expectations of the stock that a slightly bigger loss or slightly more sluggish growth didn't surprise anyone. But when a value stock outperformed expectations (which often they did), the market would take notice and quite quickly bid up the stock in anticipation of a turnaround.

The combined effect of these earnings surprises was that value stocks did well out of profit announcements because they beat expectations whereas growth stocks did poorly out of profit announcements because they underperformed expectations.

Another paper by Dreman and Berry, which refers to the above work can be downloaded by clicking here. This paper is hosted at the web site of the Institute of Psychology and Markets, publishers of the Journal of Behavioural Finance.

All of the evidence that I have seen appears to favour quite strongly the hypothesis advocated by Dreman and Lakonishok et al, that the value premium is driven by inflated expectations for growth stocks and the underrating of value stocks. Anyone who remembers the "Dot Com" bubble of the late 90s and the crash that followed should forever have etched in their memories the lesson that if you pay too much for high expected growth you will lose money. Attempts to reconcile this with the EMH and call the bubble and bust "rational" (yes, some revisionist authors say the bubble was all perfectly rational), and claim that value stocks are more risky than growth stocks stretch the limits of credulity.

I'm not fond of the pejorative term "ivory tower academic", but I think it applies very well to anybody that would argue that the Dot Com stocks which sold at thousands of times their projected sales (few had any earnings), which had few assets, undeveloped and untried business models, lost money and burnt cash at a rate faster than you can say "rights issue" and eventually went bankrupt in droves were less risky than the stalwart dividend paying "value companies" which dominate the value index.

Investing in value stocks

So how do you go about investing in a way that captures the value premium?

In a previous article, I demonstrated that due to their high costs and competitive pressures within the industry that the majority of actively managed funds fail to add value above their benchmark. There are a number of managed funds that use a value style though and some of these have beaten the major indexes like the All Ordinaries. Does this mean you should use actively managed value funds?

Not necessarily. One can still exploit the value premium without having to use actively managed funds, there are passive ways to invest in the value premium. Dimensional Fund Advisors are a passive manager that offer funds which attempt to capture the value and small cap premiums. The Fama/French indexes I have used in this article so far are the same ones that Dimensional try to replicate with their funds.

I wouldn't go so far as to call Dimensional a true "index" manager, they believe that there are ways to add value without taking on extra risk, at the expense of not being able to maintain a low "tracking error".

The two main ways in which Dimensional (and some other "index" and "enhanced index") managers try to add value is through tax management (trying to reduce the amount of realised capital gains tax distributions and hence increase the tax efficiency of the fund to a taxable investor) and through trying to minimise transaction costs.

Early index funds which were managed on a 100% non-discretionary basis often fell victim to a liquidity trap set by active fund managers. The active managers were able to anticipate changes to indexes ahead of time and, knowing that the index funds would be forced buyers or sellers on that day, would raise their asking prices to force up prices on stocks entering the index or lower their bid prices to acquire stocks at low prices that were about to be reduced in the index, knowing that the index managers would not have the discretion to avoid paying, or receiving, unfavourable prices.

It wasn't too long before index fund managers caught on to this trick, so most index funds are managed with a small amount of discretion, especially with regard to the timing of portfolio adjustments, in order to enhance returns for investors by dealing at times when conditions are more favourable.

Dimensional Fund Advisors are one of the most aggressive of all passive managers in trying to reduce their transaction costs. Using a tactic known as "block trading" they set themselves up as a market maker in many shares, especially small companies, and will offer to buy large blocks of shares at a discount to the market price or sell large blocks of shares at a premium to the market price. Other managers are happy to deal with Dimensional in this respect because it enables them to acquire or get rid of a large block of shares in a short period of time without affecting the market. Dimensional are happy to do this because they are being paid, sometimes handsomely, to provide liquidity.

As a result, Dimensional's small cap strategies especially have historically achieved negative transaction costs and have not only recouped their fees but actually outperformed appropriate index benchmarks after all costs.

But is the BtM definition of value, which is based on a company's book value, really an appropriate benchmark to use to identify value? Earlier in this article I linked to an article at the Dimensional site where they argued that BtM was superior for their purposes to other measures of value like price to earnings ratios or price to cash flow. But there are those who dispute the idea that simple quantitative benchmarks really define a "value stock".

Value from the point of view of an active investor

Warren Buffett's business partner, Charlie Munger, is scathing of the notion that value and growth are mutually exclusive. To an active value manager, a company with good prospects and a superior business is worth more than a company with poor prospects and an inferior business and hence deserves a higher price. This could be analogous to rating a painting as "good value" or "poor value" depending on its price relative to the cost of canvas, paint and frames, without taking into account whether the artist was one of the great masters or an amateur taking their first art class. Value and growth are just the opposite sides of the same coin, high growth companies are worth more and can represent "good value" even when trading at relatively high prices.

Charlie Munger had this to say at the 2000 Berkshire Hathaway annual meeting:

The whole concept of dividing it up into "value" and "growth" strikes me as twaddle. It's convenient for a bunch of pension fund consultants to get fees prattling about and a way for one advisor to distinguish himself from another. But, to me, all intelligent investing is value investing. That's a very simple concept. And I don't see how anybody could really argue with it. Buffett says, In our opinion, the two approaches are joined at the hip: Growth is always a component in the calculation of value, constituting a variable whose importance can range from negligible to enormous and whose impact can be negative as well as positive.

I completely agree with Munger and Buffett's comment. A virtually worthless company with dismal prospects can not be compared with a company with a valuable franchise and competitive advantages that would enable it to maintain a high profit margin. To put them side by side and use the same one dimensional yardstick of value would be misleading.

I agree with Buffett and Munger's comments but feel the main problem is that Buffett's definition of value is not the same as the quantitative definition of value. The two groups (active investors vs. academia) are using the same term "value" to describe something very different.

When Warren Buffett, or most active "value investors" talk about a company representing good value, they mean it is trading at a price below the price at which a businessman would be willing to purchase the entire business. They are talking about the "intrinsic value" and the concept that a company can trade at a price which is different to its intrinsic value due to market inefficiency.

Active investors definitions of value imply that one can have knowledge about a company's future earnings capacity. They arrive at a value by discounting the future cash flows into a "present value". (As a matter of fact, when I describe my process of investing in direct stocks in the article after this one, I am using Warren Buffett's definition of value and I use a similar cash flow discounting approach).

When academics refer to a "value stock", they refer to a stock that trades on low multiples of book value, earnings etc. There is nothing wrong with studying stocks on that basis, research does now confirm that if you buy such stocks you should do better than if you buy stocks trading at high multiples.

I've heard of investors who refer to themselves as "low PE" (PE = price to earnings) investors rather than value investors. Since quantitative definitions of value use a variety of different multiples, I think a more appropriate term would be to refer to the quantitative style as "low multiple" investing, or something along those lines. Systematically buying diversified portfolios of stocks trading at low multiples of assets or earnings is an approach which is quite distinct from the way most active value managers work.

Warren Buffett grew his enormous fortune not by systematically purchasing cheap low multiple stocks trading at extremely low prices, he achieved his fortune in a highly discretionary manner purchasing companies based on his estimates of value taking into account the strength of the business and future profits. Sometimes Buffett couldn't find attractive companies at attractive prices so he built up his cash positions or waited out the bull market as an arbitrageur and then if the market crashed he swooped in and bought large positions in growth stocks like Coca Cola when they were going very cheaply. Buffett has a system, but his approach isn't systematic and quantitative the way academic value approaches are.

Various studies referred to above have shown that one can't simply go extrapolating high historical growth rates into the future and expecting them to continue, it doesn't work. Indeed it shows that the profits of most fast growing companies follow a more random path and frequently revert to lower, more average, rate of growth over time.

Buffett's success implies that at least for some companies it is possible to predict their future profits with at least a small measure of certainty, but long term students of Warren Buffett know that he places enormous emphasis on understanding businesses in minute detail, in sticking to areas where you have a real competitive advantage in understanding the business, and in looking for companies with strong competitive advantages that would enable them to continue to achieve high returns despite the best efforts of their competitors.

Buffett has often said there are only a small number of companies that exhibit the kind of competitive advantages he looks for, real "strong franchises" as he puts them. Most companies lack the kind of competitive advantages that give them the market dominance required to consistently grow their profits. Given Buffett's explanation that the majority of companies can not control their markets well enough to grow their earnings steadily and predictably, there is no contradiction in Buffett's success with a handful of careful stock selections and the statistical results that say when you look at the whole market most companies' profits follow a random walk.

Active fund managers usually describe their stock picking process as one of these three approaches:

1. Growth investing
2. Growth at a reasonable price (GARP) investing
3. Value investing

Warren Buffett thinks company profit growth, quality and management are just as important as price, so although he calls himself a value investor he would be classified as a GARP investor if he ran a managed fund. As Buffett's partner, Charlie Munger said, "all intelligent investing is value investing."

But don't all fund managers, including "growth" managers exercise discipline with valuations, refusing to pay exorbitant prices for highly rated stocks? Some do, but many don't. The tech bubble provided confirmation that in order to boost their performance many fund managers were willing to load up on what can only be described as overrated speculative junk. Many growth managers both in Australia and overseas threw valuation discipline out the window near the peak of the bull market and were willing to coast along on pure momentum. The fact we even have a distinct category of fund manager "growth at a reasonable price" speaks volumes.

How have value and GARP funds performed historically?

As stated previously, only a minority of actively managed funds have beaten their market indexes in Australia and overseas. To the extent that many of the market beaters are value funds, it is appropriate to benchmark some of them against value indexes against which an even smaller number of managers outperform.

But I'm not an efficient markets hypothesis believer, I believe that there is potential for an active investor to outperform, even to outperform a value index. I wrote several articles on selecting funds in my FAQ, in particular my article in the managed funds section on Choosing a Good Active Fund.

Having said that, if you want to ride on the coat tails of the "value premium" (as academically defined), actively managed funds aren't always the best way to do it. Few active funds buy the really low priced stocks in the value indexes, at least in any great numbers. With a passively managed "value" index you at least know what you are getting.

Whenever you invest in an actively managed fund your returns will be subject to four (not three) factors (referring to the Fama/French Three Factor Model):

1. The allocation to stocks vs. bonds and cash
2. The capitalisation of the stocks you are buying, whether the fund focuses on large or small cap stocks or a blend
3. The price multiples of the stocks, as value is defined quantitatively; and
4. The skill of the manager minus their fees

A highly skilled manager will add extra returns, a highly unskilled or very expensive manager will subtract. Using the Fama/French Three Factor Model one can account for a portion of their returns, for example in 2003 small cap stocks substantially outperformed the general market and therefore small cap managers also did very well.

In the late 1990s for a few years all the "growth" managers did very well, (even though they more broadly define "growth" than just buying expensive stocks, they still did on the whole tend to purchase those expensive stocks) and then from 2000 to 2003 as value outperformed growth many of the "value" managers did quite well.

But how well did these managers perform relative to value and growth benchmarks?

There are several "value" and "growth" indexes that track the Australian market. The most commonly used ones are the S&P/Citigroup Broad Market BMI Value Index and S&P/Citigroup Broad Market BMI Growth Index. In the following chart, I have plotted the two Citigroup indexes against Dimensional's high BtM "value" index, and the ASX500 All Ordinaries Accumulation index. It is no accident that the All Ordinaries delivers a very similar performance and a high correlation with the growth index, market indexes do tend to be dominated by "growth" stocks and hence become defacto "growth" indexes. This is why nobody saw the need to invent a "growth" index fund for the Australian market, we basically already have one with the normal indexes.

Why are indexes dominated by growth stocks? Well obviously if a company goes up in price (toward a higher price to earnings ratio) its market capitalisation will increase. Since market indexes are weighted by capitalisation the stocks with the highest prices tend to dominate the index.

This chart reveals quickly why it is that I particularly hate to hear managers impose "risk budgets" where they will not allow a fund's performance to deviate more than a set percentage away from the All Ordinaries. This may not be such a problem for growth managers because growth stocks (especially large caps) perform quite similarly to the market averages. Value, on the other hand, has a lower correlation with the market index and hence is likely to be quite hampered by constraints on "tracking error". When such restrictions are imposed, in effect value managers are being told to go out and find value stocks as long as they still pad out the portfolio with lots of growth stocks.

(NB: the charts are quite similar for the US and other major markets, "large company" indexes are usually a close relative of "large company growth" indexes. Smaller company and mid-cap indexes tend to have lower average prices and hence have slightly higher correlations with the value index.)

Given the poor numbers for active investors as a whole, arguably the best way to achieve a "value" portfolio would be to simply invest in a passive value index. If you do believe you can choose a good active manager you may wish to consider augmenting your passive value exposure with some of the other "flavours" of value. Is a GARP manager a value manager? If that manager is observing good discipline in not overpaying for growth then arguably this is value investing (in the same sense as what Warren Buffett means when he says "value investing"). If your concern is that you aren't sure that the quantitative "value premium" will continue, you could at least construct a portfolio that excludes "growth at an unreasonable price" by constructing a portfolio out of passive value and GARP funds.

A portfolio which has passive and active value and GARP managers will exclude only the highest priced "growth" stocks, i.e. the ones most prone to speculative overvaluation resulting in the formation of a bubble and subsequent crash.

I do use active funds in my portfolios, researching them using the strict criteria I have set out in my article on choosing active managers. The funds I have chosen have tended to do as well as or better than the market index, but just as importantly by excluding the highest priced and most volatile sections of the market portfolio risk has been reduced.

You will have to make up your own mind about whether you have the skills or research resources to find good active funds. A combination of a passive value fund with a normal market index fund would most likely achieve a higher return over the longer term, especially after fees and taxes, than the majority of portfolios.

Emerging markets

Burton Malkiel, author of the classic book A Random Walk Down Wall Street also wrote a book called Global Bargain Hunting, which talks about the enormous profit opportunities (and risks) available to investors in the so-called "Emerging Markets". Emerging markets include Asia ex-Japan (40 years ago Japan was an emerging market), South America, Eastern Europe, the Middle East and Africa.

The economic growth in these countries is incredibly high by "western" standards, many emerging markets have had double digit rates of growth for many years, yet in terms of price earnings ratios and other quantitative value yardsticks they are cheaper.

There is no doubt that emerging markets are riskier, they are subject to regulatory and political risks far greater than those in more developed markets. On the other hand, they can also be very lucrative, as the MSCI Emerging Markets Index shows when compared to the MSCI World index in this next chart.

Emerging markets are much more risky than developed markets, as can be seen from the drawdown chart below. Since 1988 the MSCI World index has had only two bear markets where losses were greater than 20%, emerging markets have had four.

Obviously the risks are high, but so have been the returns, particularly in the early 1990s prior to the "Asian Crisis". Just as importantly, from a modern portfolio theory point of view, the correlation of emerging markets with developed markets is very low implying that some of that volatility both on the up and downside could actually result in a smoother return for a diversified portfolio that has some emerging markets rather than one that doesn't.

This chart of rolling 12 month returns gives another perspective on risk. Fortunately over most 12 month holding periods since 1988 returns have been positive.

Emerging markets are neglected by many investors who refuse to invest in them altogether. I'm not recommending investors devote a very substantial portion of their portfolios to emerging markets but it is not unreasonable for many investors to allocate a small percentage, perhaps 5% or 10% of their global shares exposure.

The major countries in the emerging markets indexes are in the "Pacific Rim" (non Japan Asia), South America and Mexico, the middle east, eastern Europe and parts of Africa. These include some of the world's most rapidly growing economies.

When people think of emerging markets as an asset class, they usually think of hyperinflation, nationalisation of assets by socialist governments or brutal military dictators, defaults on national debt, rampant corruption and a number of other rather unsavoury images. While all of these are real risks that have come up at one time or another, emerging markets as an asset class have tended despite all of this to offer superior returns to western markets, despite their high volatility. Malkiel and Mei also argue that the world really is changing as governments introduce free market reforms and increasing democracy. As these reforms take hold major risks such as revolutions and civil wars decrease.

An interesting chapter of Global Bargain Hunting contains the following passage, which shows how far emerging markets can come in a short time:

At the beginning of the 1990s, a China Products Fair would bring a wry smile to the faces of browsing foreign traders. A story in the Nikkei Weekly described these products as "the clunky toaster; the cheesy-looking welder; a tank-sized refrigerator looking as if it belonged at the head of a May Day parade. At best functional, always 20 years out of date. Chinese products were destined almost exclusively for Third World markets."

There have certainly been some improvements there, if you go down to your local electronics retailer you'll find Chinese high technology items like DVD players, plasma televisions and other consumer goods, of not much different quality to the Japanese goods next to them (but usually at much lower prices). The Chinese aren't just exporting these goods, a growing middle class is consuming it also. Car sales are sky-rocketing in China, enormous cities are being built in China's east to cater to incoming peasants from the western regions looking for work in the factories and this is bringing boom times to other industries, most notably the suppliers of raw materials like steel and concrete.

The same is going on in Eastern Europe, where major manufacturers are setting up factories in Poland, Hungary and other former Soviet Union members. In South America also, Argentina, Brazil, Chile and Mexico, for many years economic basket cases that went from crisis to crisis, are now capitalist democracies and their economies have responded accordingly.

The rate of growth in per capita income in many emerging markets has not been seen in currently developed economies since the post WW2 re-emergence of Japan and the great boom in the United States in the 19th century (note, at the time of these booms, both Japan and the USA were considered emerging markets). With this extra wealth has come extra consumption and a maturing of these markets. The largest holdings in emerging markets index funds (like those offered by Vanguard and DFA) are major telecommunications carriers in Asia and South America, oil refineries and some of the world's biggest manufacturers of consumer goods. The largest company in the emerging markets index is none other than South Korea's Samsung, you may recognise other companies in this portfolio as well.

For the most part you could say that the biggest problem with Emerging Markets companies is not the companies themselves but where they are listed. In the book, Malkiel and Mei repeatedly make the point that company growth rates in emerging markets are on average much higher than those in developed markets but price to earnings ratios are much lower. The authors argue that emerging markets represent a great bargain for those willing to stick with them for the long term.

Downloading a copy of Vanguard's fact sheet for their Emerging Markets Shares index fund (only available to wholesale investors with $1,000,000 to invest, but accessible for the rest of us via a wrap account), many of the top ten companies are, if not household names, at least recognisable.

1. Petroleo Brasileiro (Brazil)
2. Gazprom (Russia)
3. Cia Vale Do Rio Doce (Brazil)
4. China Mobile (China)
5. Samsung Electronics (Korea)
6. America Movil (Mexico)
7. Taiwan Semiconductor (Taiwan)
8. Lukoil (Russia)
9. Reliance Industries (India)
10. Posco (Korea)

Source: Vanguard INDEX FUND FACT SHEET - 31 March 2008

So we aren't talking about yak herding companies here, many of the world's largest semiconductor manufacturers (like Taiwan Semiconductor) and some very large general manufacturers (like Samsung) are listed in emerging markets. China Mobile is another big name with a dominant position in the rapidly growing Chinese mobile phone market.  Gazprom and Lukoil of course will be well known to anyone who watches the news and is aware of events in Russia.

Literacy rates in some of these countries are as high as, or higher than, many countries in the western world, and the work ethics of many (most notoriously the South Koreans) are legendary. Democracy is also spreading, with countries like Chile and Argentina finally enjoying representative government.

Socialism has been rolling back and even India, formerly an ineptly managed centrally planned socialist economy has been awakened by their great rival to the north (the ostensibly communist China) and themselves are embracing market reforms and trying to tackle corruption.

Emerging markets are certainly not a safe asset for short holding periods but for a patient investor with a lot of time on their hands the potential rewards are tremendous.

Malkiel and Mei argue strongly in Global Bargain Hunting that passive or indexed approaches are the best way to invest in emerging markets, despite the supposed inefficiency in these markets. The reasoning behind this, which is backed up with a significant amount of data, is that transaction costs are very high in these markets and portfolio turnover can be extremely expensive. Their evidence shows that whatever benefit active investors might gain from active investing tends to be swallowed up by the high costs of trading on emerging markets.

Value and small company premiums have also been demonstrated to exist in emerging markets and although Dimensional Fund Advisors offer emerging markets value and small company funds to American investors, in Australia so far there is only a non-value emerging markets fund.

Momentum

"Momentum" is an anomaly that Efficient Market Hypothesis' true believers particularly dislike. If there is momentum in stock prices then it means you can achieve above average performance simply by buying stocks that are going up a lot. You just look at a chart of stock prices, if it has significantly outperformed the market in the last few months you buy it, if it has underperformed the market you sell it. This is a favourite strategy for professional and amateur investors alike, and is particularly loved by people that prefer technical analysis (chart reading) over fundamental analysis (assessment of the economics of the business).

What evidence is there for momentum? There have been a number of studies done, and they have found that in the short term there is indeed a "momentum premium". For example James P. O'Shaughnessey studied momentum for his book What Works on Wall Street (another book I summarised in my FAQ), and he found that the strategy of buying stocks with the highest performance in the last year did lead to high performance this year as well, furthermore he found that portfolios formed from last year's worst performing stocks would underperform this year as well.

Another study, by Louis Chan, Narasimhan Jegadeesh and Josef Lakonishok (The Journal of Finance 51 (no. 5), December 1996) found that portfolios created with the highest price momentum and earnings momentum from the past six months did lead to higher performance over the next six months, twelve months, twenty four months and even thirty six months. There was a fairly clean linear relationship, the higher past momentum portfolios performed better than the lower past momentum portfolios.

While the above is not in any real doubt, we have to ask ourselves if this is a good investment strategy. One thing that is very clear is that momentum chasing leads to a very high portfolio turnover. You need to buy and sell a lot to keep your portfolio always stocked up with the highest momentum stocks. In my article "why invest long term" I explained that this leads to substantial expenses, in particular higher taxes. I am unconvinced that the short term "momentum premium" is really exploitable, although these researchers have found that high momentum portfolios beat neutral momentum portfolios (i.e. index funds) that the higher turnover would probably wipe out the extra gains for a taxable investor.

These momentum studies dealt with short term recent performance, O'Shaughnessy was looking at portfolios formed from the stocks with the best 12 month price performance, Chan et al were looking at portfolios formed based on past six month performance. There have been studies based on past three year and five year performance, and they have come to dramatically different conclusions regarding the wisdom of chasing high past performance.

One of the better known examples was a paper by Werner F.M. DeBondt and Richard Thaler, Professors at the University of Wisconsin and Cornell University, respectively. They examined the investment performance of stocks with the worst and best prior investment results in "Does the Stock Market Overreact?", The Journal of Finance, July, 1985.

Rather than one year momentum, they looked at the subsequent performance of portfolios formed out of the 35 worst and 35 best performing companies selected each year based on previous five year performance, reforming the portfolio annually on 31 December from 1932 to 1977.

Their conclusions were different to conclusions reached by the other studies mentioned above. They found that based on five year momentum you are better off buying underperformers than outperformers.

The 35 stock "dog" portfolio outperformed the market benchmark by 12.2%pa, compounded, and the 35 stock "glamour" portfolio underperformed the market by 4.3%pa.

The study was repeated for stocks based on three year past performance and the results were similar. You are clearly better off buying long term bad performers than outperformers, though if you reconcile this with the other research it seems you could do well by buying stocks with a rotten five year past performance and a decent one year past performance.

This is interesting, because it means that if you are looking at longer term past performance you seem to be better off sticking with underperforming stocks, because they tend to have further to rise than the stocks with the greatest past appreciation. Many of the stocks that have taken a beating over many years are "value" stocks, the market has very low expectations for them and hence they have been sold down to very cheap prices.

David Dreman found in his work on earnings surprises that analysts were slow to adjust their assessments of companies. Analysts systematically underestimated the performance of value companies, and when they announced profits exceeding analysts expectations analysts took their time to readjust forecasts.

Every time a company announces a profit greater or less than analyst's consensus estimates, the stock price will move in response. As I mentioned above when writing about value vs. growth stocks, value stocks tend to benefit from earnings surprises, but on average growth stocks suffered. Dreman found that it sometimes takes several years for analysts to fully revise their expectations about a company and so companies tend to string together a number of earnings surprises over several years. This leads to sustained outperformance from value stocks, and sustained underperformance from growth stocks.

If you combine each of these pieces of data together you can start to see the bigger picture. Value stocks usually become value stocks because analysts don't see much hope of the company growing their profits over time. Because analysts feel this way about these companies the stocks tend to underperform the market significantly for a number of years, at the end of this period of underperformance the stock will usually be a fully fledged value stock, with an excellent dividend yield, low price to book value ratio, low price to earnings and other measures of good value.

Changes in the company or the market leads to a value company announcing higher than expected profits. This doesn't make it an exciting buy though, because the market wasn't expecting much. Nevertheless this earnings surprise leads to a jump in the price over the course of a few months. As the next few quarters come and go it becomes apparent that the company is fundamentally changing, and analysts are forced to revise their forecasts upwards. This leads to a series of ratings upgrades over an extended period of time, and sustained high performance.

If this is true, then one would think that the best performing strategy to use would be to look for stocks with poor long term performance, trading at low prices, that has just surprised the market with a higher than expected profit. I haven't seen a study yet that ties together all of these things, but it is interesting that in What Works on Wall Street the highest performing strategies he found combined value measures (such as low price to earnings ratios and in particular low price to sales ratios) with momentum (highest performance of all value stocks over the last 12 months).

If you focus your attention on value stocks you can hold the stock for a much longer period of time than someone chasing momentum without applying a value filter. While overall stocks with high momentum do seem to outperform over the next year, the outperformance diminishes rapidly past the first twelve months. Stocks that exhibit both value and momentum tend to outperform for a much longer period of time, so you can hold these stocks for several years instead of just one, thus reducing your turnover substantially.

Momentum is in fact used successfully by some supposedly "passive" fund managers.  Dimensional Fund Advisors, who normally assert that markets are quite efficient, nevertheless incorporate various momentum filters into their buying and selling decisions, designed to delay purchases on stocks which are still significantly underperforming the market, and hold off selling while a stock is still going up.  While there is no ready efficient market explanation for why this works, the fact that it does work is enough to justify using it. 

Market efficiency and the index strategy

If a market is highly efficient, then it means information is quickly built in to prices and therefore there is no such thing as an undervalued stock. If there is no such thing as an undervalued stock, then there are no strategies that could lead to consistently high performance over the long term. The evidence given above is that the market is not quite fully efficient. There are a very small number of managers that have demonstrated an ability to perform well that goes beyond mere luck. Low priced "value" stocks do outperform more expensive stocks. What does all this mean to indexing?

As I made quite clear in the first article on managed funds, no matter what a group of investors do, half will always be below average. If there are a great many highly skilled investors then it would be fair to say that underpriced stocks are quickly bought up by bargain hunters until this underpricing is gone. Markets are not totally efficient, but they are pretty close to it most of the time.

Indexing is not an "efficient markets" strategy. Obviously even if markets were grossly inefficient and individual stocks were very under-priced or overpriced, index funds would still come out about average and would beat half of all active funds. Since the costs of index management are necessarily going to be significantly lower than active management it will always be a feature of markets, efficient or not, that index funds will beat most managed funds. In addition, because of competition in the funds management industry it would be fair to say that advantages enjoyed by the top managed funds won't persist for long, and regression to the mean will kick in eventually.

There is a middle ground between market indexing and active management. Active managers try to take advantage of known inefficiencies yet incur large costs in doing so. Passive approaches with a value and small cap emphasis offer, we believe, the most realistic chance of delivering the value and small cap premiums to investors because they don't squander returns through excessive costs and because they can be designed with capturing those premiums exclusively in mind.

---