Significance of the Study Modern Finance Theory
Modern Finance Theory paradigm created over 50 years ago, of which the EMH is the cornerstone, would be challenged and this would encourage future research in Behavioral Finance and strengthen the case for a paradigm shift in price formation theory.
Empirical evidence of market inefficiency and consequent inability of the market to manage prices and risk, will strengthen the case of government bureaucrats pushing for more stringent global regulatory standards on the necessary level of bank capital, liquidity and leverage requirements and overall banking industry regulation in the aftermath of the 2008 global financial crisis.
Allocation of Resources
Inefficient capital markets will send inaccurate signals to private and institutional investors regarding their capital investment decisions amongst the nation’s private and public companies, and firms will receive inaccurate price signals to guide their strategic production and investment decisions. This will result in the sub-optimal allocation of the nation’s resources in unproductive companies, which will have a significant detrimental long term effect on economic growth and prosperity.
Long-Term Investment Strategy
From a long term investment strategy perspective, global capital market inefficiency dictates that it would be a wise choice to select a professional investment manager who is able to capitalize on global market inefficiencies by diversifying and trading internationally, thereby generating returns in excess of the market and a much needed larger nest egg upon retirement.
The EMH as the cornerstone of Modern Finance Theory and technical analysis as an intensively and widely used investment analysis technique, together constitute vital elements of Financial Management which remains a core foundation course of all Master of Business Administration (MBA) programs2. In addition, top business schools place more than 40% of their MBA graduates in the finance industry – former Merrill Lynch CEO’s John Thain and Stanley O’Neal, Lehman Brothers CEO Richard Fuld, and ex Citigroup CEO Vikram Pandit are all card-carrying MBA graduates (Holland, 2009).
A disturbing excess volatility anomaly in stock market prices was uncovered by Shiller (1981) who found that stock market price changes over the past century were statistically five to thirteen times too high to be attributed to subsequent new dividend information. This implies that price changes occur due to unexplained reasons other than subsequent new information, which calls into question the entire underpinnings of the EMH.
The “Impossibility” of the EMH
Grossman and Stiglitz (1980) argued that it is impossible for markets to be informationally efficient in the real world where information is costly: if prices fully reflected all available information, informed traders who expended resources to obtain that information would not earn a return and be compensated for their efforts. An unrealistic theoretical world of costless information was therefore a necessary condition for efficient markets but even then the markets would not be in equilibrium as traders would realize that additional profits could be earned by expending resources and becoming more informed. Fama (1991) later agreed with this view and proposed an economically more sensible version of the EMH which stated that prices reflect information to the point where the marginal profits of trading that information do not exceed the marginal costs of acquiring it.
In addition, Oppenheimer and Schlarbaum (1981) found strong contradictory evidence against the EMH by showing that positive abnormal risk-adjusted rates of return were available to investors who followed the value based fundamental analysis stock selection criteria developed by Columbia University professor Benjamin Graham4. More salt was rubbed in the EMH wound created by the contradictory size, value, seasonal and excess volatility evidence with the groundbreaking paper by De Bondt and Thaler (1985). Their findings heralded the beginning of Behavioral Finance, which provides a theoretical challenge to the EMH by questioning the rationality assumption of investors underpinning the EMH and suggesting that cognitive psychology better describes investor behavior. De Bondt and Thaler postulated that traders are human beings with feelings and emotions which result in irrational trading decisions such as the overreaction to market news. This overreaction results in price trends and in extreme cases even market bubbles and crashes. The hypothesis was tested that traders view past winners (stocks that have outperformed the market) as overpriced and past losers (stocks that have underperformed the market) as underpriced, and then basically overreact to this information. Their findings supported the overreaction hypothesis with a portfolio of prior “losers” outperforming the market by 19.6% and a portfolio of prior “winners” underperforming the market by 5% in the 36 months following portfolio formation. In addition, their findings supported the seasonal “January effect” with most of the excess returns being realized in the month of January.
Motivated by the work of DeBondt and Thaler on the overreaction effect and its potential impact on contrarian trading strategies, Jegadeesh and Titman (1993) decided to study the impact of over and under reaction to the alternative relative strength strategy of buying past winners and selling past losers. Their findings indicated initial significantly positive returns in the first 12 months, followed by substantially negative returns thereafter in years 2 and 3. This was the first evidence of short-term momentum and long-term mean reversion in stock price time series, a serious blow to the EMH
At the time, the lone island in this sea of contradictory size, value, seasonal and behavioral finance evidence was Eun and Shim (1989) who managed to find support for the EMH by using vector autoregressive analysis to show that the relative speed at which international markets reacted to US market shocks proved that international stock markets were informationally efficient.
Shortly thereafter in his second review paper, Fama (1991) attempted to defend the EMH by positing the joint hypothesis problem: tests of the EMH are in fact not only testing market efficiency but also the equilibrium asset pricing model used to statistically analyze the findings. A rejection of the EMH could therefore actually mean a rejection of a misspecified asset pricing model or both. This became known as the “bad model” problem (Fama, 1991). Fama also suggested that the seasonality anomalies may be due to data mining and errors in the Center for Research in Security Prices (CRSP) historical data base utilized in all the seasonality studies. Later in his third review paper, Fama (1998) made the related point that most anomalies were the result of the research methodology used and disappeared with a change in technique. This view was supported by Malkiel (2003) who added that transaction costs would also offset any potential profits that traders could actively glean from the uncovered anomalies. However, Jegadeesh and Titman (2001) argued that their later study using nine additional years of out-of-sample data still showed short term momentum and the persistence of this anomaly.
Fama (1998) also defended the EMH against the Behavioral Finance findings of De Bondt and Thaler (1985) and Jegadeesh and Titman (1993) by positing that overreaction by market participants occurred as frequently as underreaction, and this random split of overreaction versus underreaction was actually consistent with market efficiency.
Atkins and Dyl (1993) also challenged the outperformance due to the overreaction effect postulated by De Bondt and Thaler (1985) on the basis that previous overreaction studies assumed that traders could buy and sell at the database closing prices. Atkins and Dyl argued that in reality traders had to buy at the quoted ask price and sell at the quoted bid price, and posited that the bid-ask spread would have a detrimental effect on earnings of the overreaction strategy of buying past losers and selling past winners. After taking the bid-ask spread into account, Atkins and Dyl (1993: 95) indeed found that the cumulative abnormal returns generated by the overreaction strategy were less than the bid-ask spread, thereby leading them to conclude that contrary to previous overreaction effect studies, no profitableopportunities were present due to the size of the bid-ask spread and “no compelling evidence of market inefficiency is found”.
The Weak-Form EMH proposes that security prices at any point in time reflect all past historical information. Security prices therefore have no memory and the best estimate of a future price is the current price.
The weak form EMH therefore implies that technical analysis (which relies on charts and analysis of past price patterns to predict future price movements) would be of no value and cannot generate returns in excess of the market.
Following the literature review of prior research which all supported the Weak-Form EMH, technical analysis should prove to be a worthless activity, yet technical analysis has endured over time and is still an intensively and widely used investment analysis technique. In fact, in an extensive survey of 200 foreign exchange and international fund managers, Gehrig and Menkhoff (2006) found that technical analysts make up the largest group among foreign exchange traders and the second largest group among international fund managers.
This indicates a clear disconnect between technical analysis as employed by practitioners in the market and the technical analysis methodologies utilized by academics in prior Weak-Form EMH studies.
The problem is that all previous technical analysis EMH research neglected to exploit inter-market technical analysis, the powerful combination of qualitative and quantitative techniques, high frequency intra-day strategies and volume confirmation signals in addition to focusing exclusively on individual domestic stock or currency markets.
These issues cumulatively result in methodological weaknesses which severely handicap the profit generating potential of technical analysis and suggest that previous Weak-Form EMH research findings were erroneous in being unable to reject the null Weak-Form market efficiency hypothesis.
This study therefore proposes that by eliminating prior methodological weaknesses and utilizing high frequency intra-day data, the combination of qualitative and quantitative techniques and volume signals to develop intermarket technical analysis strategies, it is possible to generate significant excess profits and consequently show that contrary to prior research findings, the developed country capital markets are not Weak-Form efficient.
Trading the Precious Metals
Intermarket relationships as identified by Ruggiero (1997) were utilized to develop intermarket momentum trading strategies for the Precious Metals – Gold (GC), Silver (SI) and Platinum (PL).
The Trade Strategy Premise
Ruggiero (1997: 21) illustrates with the aid of graphs the intermarket relationship that clearly shows the Gold price being negatively correlated with the US Dollar due to the previously described international reserve currency characteristic of the US Dollar.
It follows that Gold will also therefore be positively correlated to Commodities. This can also be elucidated by the fact that Commodities form the basic inputs that fuel the engine of an economy, therefore a rise in commodity prices will generally flow through the production chain and result in increased producer and consumer prices. Heightened inflationary expectations increase the demand for the store of value inflation hedge characteristic of gold. This chain of events initiated by a rise in Commodity prices results in an appreciating gold price reinforcing the positive relationship between the Gold price and Commodities.
It was previously shown how a positive relationship between Commodities and US Treasury Bonds may be deduced from the positive correlation between the Commodity Currencies and US Treasury Bonds. Therefore, one can further deduce an expected positive correlation between Gold and US Treasury Bonds which can also be clarified by the link between rising bond prices and the store of value inflation hedge characteristic of Gold. As bond prices rise, bond yields and interest rates fall – in a low interest rate environment, companies prosper thus raising future inflation expectations which consequently stimulate the demand for Gold as an inflation hedge.
The nature of Gold mining stocks to predict and lead the Gold price at major turning points was explained by Ruggiero (1997) as arising from the fact that Gold stocks are highly leveraged to the price of Gold. For example, a 9% rise in the Gold price from $ 1100 to $ 1200 per ounce will result in a 100% increase in the profit of a Gold mining company with a production cost of $ 1000. This predictive positive correlation was confirmed by charts depicting the price of Gold versus the Philadelphia Gold and Silver Index (XAU) – a capitalization weighted index of the 16 leading precious metal mining companies involved in the mining and production of gold and silver (Ruggiero, 1997: 24).
Finally, Ruggiero (1997: 23) illustrates the close correlation between the Precious Metals – Gold, Silver and Platinum, with Silver and Platinum leading and being predictive of Gold at major turning points. This correlation is due to the fact that apart from their industrial usage, the Precious Metals – Gold, Silver and Platinum, more importantly all derive value due to their function as hedges against inflation and a store of value. Their market prices therefore not surprisingly move together very closely, and the above noted intermarket relationships for Gold will also therefore apply to Silver and Platinum.
Number of Trades
As detailed in Table 6-7, the Intermarket Momentum portfolio generated 173 trades over the entire data sample period from January 1, 2008 to December 31, 2013 with the trades evenly distributed between the Test set (January 1, 2008 to December 31, 2012) and the Live set (January 1, 2013 to December 31, 2013).
Trend following momentum based strategies are dependent upon market volatility to generate the substantial price movements required to trigger trades. The relatively large number of trades experienced during the Live set is testimony to the high market volatility in 2013 resulting from the Federal Reserve’s unprecedented $ 1 trillion a year quantitative easing (QE) bond buying program, the market’s continuing speculation as to whether the Federal Reserve will begin to taper QE or not, and the Bank of Japan (BOJ) following the lead of their US counterparts by embarking on a similar groundbreaking $ 70 billion a month bond purchasing program in an effort to reinvigorate the Japanese economy.
The single trade outlier in the year 2009 was due to the low market volatility experienced in the aftermath of the 2008 Financial Market Crisis. Removing that single trade outlier from the analysis generated an average of 34 trades a year, or approximately 3 trades a month which is of sufficient frequency to keep the trader engaged. Prolonged flat periods out of the market, such as that experienced in 2009, severely test the discipline and patience of the trader to stay the course and remain true to the strategy.
In order to be profitable, a strategy requires either a high winning percentage above 50% or a high average win to average loss ratio. The Intermarket Momentum portfolio had a low 38.15% winning percentage but the average winning trade ($ 102 956.52) was more than double the average losing trade ($ 46 681.10) resulting in its overall profitability with a 2.21 average win : average loss ratio. The portfolio also traded on both the long and the short side, with the short trades slightly more profitable (44.16%) than the long trades (33.33%).
As an indication of a strategy’s robustness and ability to trade profitably in the future, there should be no significant deterioration in performance from the Test set to the Live set. The Intermarket Momentum portfolio performed remarkably well in the Live set, almost matching the Test set winning percentage whilst exceeding the Test set winning percentage on long trades, and experiencing a positive reduction in the size of the average losing trade. There was however a slight deterioration in the size of the average trade net profit, the winning percentage on short trades and the average winning trade.
To avoid the risk of ruin, the maximum drawdown (defined as the maximum peak to trough decline in trading capital measured as a percentage) should not exceed the threshold of 40% of initial capital. The Intermarket Momentum portfolio experienced a $ 533 753.50 drawdown, equivalent to 53.38% of the initial capital, within the first three quarters of the data sample period between January 1 and Augus 7,2008.
This drawdown suggests that a prudent level of starting capital to trade the portfolio with a trade size equal to 10% of capital, would be $ 1.5 million resulting in a much safer 36% maximum drawdown.
Hypothesis Testing The Null Hypothesis
The Null Hypothesis supports the Weak-Form EMH by stating, given all publicly available historical information, that an alpha less than or equal to zero indicates a portfolio of Intermarket strategies cannot generate returns in excess of the market, thereby implying that the developed country capital markets are Weak-Form efficient:
The Alternative Hypothesis
The Alternative Hypothesis states that a significantly positive Alpha α indicates risk-adjusted performance in excess of the market which implies that the developed country capital markets are not Weak-Form efficient:
The positive Alpha of 8.52% is > 0 and significant at the 5% level (t-statistic of 3.90 > the 1.96 critical value). This indicates that the positive Alpha generated by the Intermarket Momentum Portfolio is unlikely to have occurred by chance alone and that there is less than a 5% probability of erroneously rejecting the Null Hypothesis.
This allows the rejection of the Null Hypothesis and the acceptance of the Alternative Hypothesis that the developed country capital markets are not Weak-Form efficient.
The trade by trade details generated by the TradeStation strategy performance reports enabled the synthesis of aggregate summary performance statistics over the entire data sample period from January 1, 2008 to December 31, 2013.
This allowed the calculation and test of significance of the Alpha generated by the portfolio of Intermarket Momentum trading strategies and the consequent testing of the Null Hypothesis.
Significance of the Study
Modern Finance Theory
Allocation of Resources
Organization of the Study
2 LITERATURE REVIEW
Strong Form EMH
The Underlying Theory
The “Fair Game” Model
The Martingale Model
The Random Walk Model
The Empirical Evidence
3 RESEARCH QUESTION
Qualitative or Quantitative Studies
Daily Closing Price Data Series
The Null Hypothesis
The Alternative Hypothesis
TradeStation Trading Platform
CSI Correlation Lab
IBM SPSS Statistics Software
High Frequency Data
Trading the Commodity Currencies
The 30 Year US Treasury Bond Trading Strategy
Number of Trades
7 CONCLUSIONS AND RECOMMENDATIONS
Significance of the Results
Modern Finance Theory
Allocation of Resources
Long-Term Investment Strategy
Recommendations for Future Research
9 GLOSSARY OF TERMS
GET THE COMPLETE PROJECT
Evidence that Weak-Form Capital Market Efficiency Does Not Hold