This relationship held through the great bull market that ran from through As rates fell, bond price rose along with stock prices, exactly as the intermarket relationship forecast. As commodities fell, their negative correlation with the dollar tells us we should expect the dollar to rise. Data for the US dollar trade-weighted index begins in and the dollar was lower in than it was when the data series began. This relationship did not hold up as predicted.
The failure of intermarket relationships to hold up over the long run led to a reevaluation of the theory and the new idea that in a deflationary environment some of the relationships will change. With the revisions to the theory, we may be getting into an exercise in curve-fitting. Under this theory there will be times when stocks and bonds move together and other times when they move in opposite directions.
The distinction is supposed to be defined by inflation expectations. This less precise relationship can be confirmed by measuring the correlations of different price series over different time periods.
Intermarket Trading Strategies: An Overview of How Other Markets Impact Stocks
The chart below shows correlations between the dollar and gold prices, the dollar and oil prices, and oil and gold prices. The relationships vary a great deal over time with correlations moving from positive to negative in an almost random fashion. Because correlations change over time, intermarket strategies are best used for short-term trading.
Correlations of different asset classes should be calculated every few months and portfolios built to benefit from the correlations. For example, adding assets with negative correlations can reduce risk. However, because correlations change the portfolio will need to be reviewed and rebalanced relatively frequently. In other words, intermarket trading strategies are not really very useful for individual investors. But intermarket investment strategies can be useful in the long run. To implement these strategies, we need to understand what different asset classes represent.
Historically, emerging market stocks have delivered the best returns but carry the greatest risk. Many emerging markets are highly correlated to natural resources because energy and mining sectors may represent a large portion of the market capitalization of these markets. He understands and explains the important principles underlying successful trading and offers his solutions. I can highly recommend it. Markos Katsanos does an excellent job of explaining the relationships between different markets through the use of statistical tools, which he does in clear and simple language.
Readers of this notable book will be better able to create trading systems that can conquer any market. Prodotti correlati. Der objektivierte Unternehmenswert. The next composite chart in Fig. Intermarket Analysis 5 Figure 1. The above examples are included to illustrate that the integration of global markets can extend beyond the obvious relations. Although some emerging markets may have relatively medium to low correlation with US markets, one important question to ask is whether diversification works when it is needed most. Evidence from stock market history suggests that periods of negative shocks and poor market performance were associated with high, rather than low, correlations.
The events of 21 January are still fresh in my mind, when a 2. Indeed, investors who have apparently relied upon diversification in the past to protect them against corrections of the market have been frequently disappointed. These can only be created by taking into consideration directional movements of correlated markets.
The use of intermarket correlation analysis can help you improve on your trading system by avoiding trades against the prevailing direction of correlated markets, but can also be used on its own to develop a complete system based on divergences between two or more highly correlated markets. Knowing the correlation of the market you propose to trade with other markets is very important for predicting its future direction. Asian markets are the first to start trading, followed by the European markets.
For a US trader the insight gained from all preceding markets is a valuable tool in predicting at least the opening in his local market. I have found that the most accurate economist is the market itself. It is far easier to forecast economic activity from the behavior of markets themselves than it is to forecast the capital markets from lagging economic statistics such as the unemployment index.
The market is a discounting mechanism. It interprets the impact of economic news some time in the future. Of course, this is only a guess and guesses are not always right. But the truth is that the market is a much better guesser than any of us are, as it represents the average opinion of all the economists in the world. There appears to be no end to the conclusions that can be drawn if a little understanding, imagination, and pure common sense are applied. Major changes in commodity prices affect the bond markets of different countries in different ways, depending upon their economic structure.
What sectors are affected first? Which asset class will provide the best potential profits? If opportunities dry up in one sector, where is the money heading to take advantage of the next cycle? This is what intermarket analysis can tell you if you learn what to look for, which makes it a grand endeavor and a continuing challenge but always worth the effort.
Intermarket analysis can also be useful in estimating the duration and state of the business cycle by watching the historic relationship between bonds, stocks and commodities as economic slowing favors bonds over stocks and commodities. Near the end of an economic expansion bonds usually turn down before stocks and commodities and the reverse is true during an economic expansion. Bonds are usually the first to peak and the first to bottom and can therefore provide ample warning of the start or the end of a recession. Bonds have an impressive record as a leading indicator for the stock market, although this information cannot be used in constructing a trading system as the lead times can be quite long, ranging from one to two years.
You can see in Fig.
The Commodity Research Bureau CRB index was the last to peak, making a complex triple top formation with the last peak coinciding with the start of the recession. Whatever the relationship is — leading, lagging, or divergent responses to economic conditions — a strong negative correlation coefficient between two markets is a suggestion that these markets will move against each other sometime in the future. And, of course, the higher the absolute value of the coefficient of correlation, the higher the diversity of their performances.
Although intermarket analysis has been classified as a branch of technical analysis, it has not been embraced fully by analysts. The majority of traders continue to focus on only one market at a time and they tend to miss the forest for the trees. No market exists in a vacuum, and traders who focus on the bigger picture portrayed through all international markets tend to be the ones that deliver better performance.
Traditional technical analysis indicators such as moving averages are lagging indicators calculated from past data and are limited in assessing the current trend. Regardless of the hours spent in back-testing, there is a limit beyond which a system based on a lagging indicator can be improved further.
Thus the addition of leading indicators that anticipate reversals in trend direction is essential and beneficial to Intermarket Analysis 1. A custom indicator can also be calculated from the ratio of prices, to help assess their past relation and anticipate future direction. Both of the above methods, however, are limited to two markets and the use of the correlation coefficient is essential for an analysis of multiple markets. For predictive purposes, we wish to detect correlations that are significantly different from zero.
Such relationships can then be used to predict the future course of events in trading systems or forecasting models. In addition, linear regression can be used to predict the future price trend of a market based on its correlation with multiple related markets.
When assessing intermarket relations you should always keep in mind that these are neither fixed nor static in time. Instead they fluctuate continuously in strength and time. It is usually very difficult to determine which market is leading or lagging. A lead can shift over time and become a lag, with the markets switching positions as follower and leader. In addition, a weak positive correlation can sometimes become negative and vice versa. For this reason it is always prudent to look at the prevailing rate of change of the correlation between two related markets before reaching any important conclusions or trading decisions.
The variability of the correlation over time is more evident in Fig. You can see that correlations before were inconsistent and unpredictable but started to converge during the last tenyear period. Less common, however, is the diversification into other asset classes such as commodities or foreign currencies forex. Notice the correlation volatility, especially before There is a widely held belief that, because commodities and currencies are traded on very thin margins, they are just too risky and can lead to financial ruin. Because of their low correlation to equities, most commodities are attractive diversification candidates as they can lead to a large increase in return while simultaneously reducing risk.
Furthermore, futures diversification is particularly effective in declining stock markets, just where it is needed most. During periods of very low or negative stock returns, commodities except industrial metal futures dominate the portfolio return, acting as a hedge, or buffer, in falling markets. The benefit of including foreign stocks is not so clear as the world has gotten smaller due to the ability to communicate almost instantaneously. Intermarket Analysis 11 Unfortunately, the approach of most novice investors or even fund managers is to have no risk management at all and it becomes obvious too late that this is an extremely dangerous omission.
A fund or portfolio manager should not be evaluated only by the return he has achieved. Another important criterion of his performance is the portfolio risk exposure over time. A good benchmark of that risk is the standard deviation of returns. This is a measure of how far apart the monthly or yearly returns are scattered around the average.
Correlation is a relatively simple concept but absolutely mandatory in the use of investments. It basically refers to whether or not different investments or asset classes will move at the same time for the same reason and in the same direction. To be effective, diversification must involve asset classes that are not correlated that is, they do not move in the same direction at the same time.
High positive correlation reduces the benefits of diversification. On the other hand, selecting uncorrelated or negatively correlated asset classes not only reduces the downside volatility in the performance curve of the portfolio to a minimum but can also increase overall profitability as well.
An example will help illustrate the basics of diversification. Suppose you are considering diversifying your stock portfolio by adding an uncorrelated commodity future from the energy complex. If you invest your entire equity in either stocks or crude oil futures, and returns vary in the future as they have in the past, your equity line in points will be similar to the charts in the bottom window of Fig. The reason for the reduction in volatility is that stocks did not move in the same direction at the same time with crude oil futures.
Thus, a crucial factor for constructing portfolios is the degree of correlation between investment returns.
Diversification provides substantial risk reduction if the components of a portfolio are uncorrelated. In fact, it is possible to reduce the overall risk of the portfolio to almost zero if enough investment opportunities having non-correlated returns are combined together! Maximum return, however, is also proportional to risk.
Intermarket Trading Strategies
Low risk investments produce low returns and speculative or riskier investments can produce higher returns. Thus reducing risk can also reduce return. Like everything else in life, the best solution is a compromise between risk and return. The following example will help illustrate the basics of selecting an appropriate portfolio of securities or asset classes.
Down arrows indicate major tops in stocks and bonds and up arrows bottoms. Bonds were also the first to bottom in anticipation of the recovery, followed by commodities and then stocks. From the beginning of until the middle of all three were rising together. Commodities are usually the last to bottom during a recovery but this was not the case here as they were boosted by the weakness in the dollar.
The dollar made a final peak in January and reversed direction, dropping like a rock against the euro and other major currencies. This triggered a secular bull market in gold which spread to the rest of the commodities and has continued until the end of June , almost nine months after stocks peaked in September The composite portfolio produced better returns with less volatility. In Table 1. The correlation coefficients between the selected asset classes or indices are listed in Table 1.
These coefficients are based on monthly percentage yields and are calculated, as part of this study, over the same year period. As discussed later, the correlation coefficients play a role in selecting the asset class allocation. You can see from Table 1. Gold is Intermarket Analysis 13 Table 1. Also foreign index returns were converted to US dollars. As you will see later, however, in the case of low correlations the volatility and not the correlation coefficient is the dominant factor to consider in reducing portfolio risk. Table 1. The standard deviation of each portfolio was calculated in a separate spreadsheet by adding the annual returns of each asset class according to their percentage weights in the portfolio and then calculating the standard deviation of the annual returns for the entire year period of the study.
The first portfolio in the third column of Table 1. Commodities, however, were the real star of the show as they played an important role in significantly reducing risk and, at the same time, increasing return. This is evident from the standard deviation of returns of the third hypothetical portfolio. The fourth portfolio was obtained by finding the best allocation highest return with the minimum risk.
The relatively high percentage allocation of bonds Intermarket Analysis 15 was to be expected as their standard deviation was the lowest of the group. The presence of the highly volatile oil futures in the minimum risk portfolio, however, was certainly a surprise. The relatively low performance of this portfolio was no surprise as the standard deviation is proportional to returns: the smaller the standard deviation, the smaller the risk and, of course, the smaller the potential magnitude of the return.
There is therefore a limit beyond which the expected return cannot be increased without increasing risk. This portfolio last column in Table 1. Similarly in minimizing risk fourth portfolio I had to specify a minimum return otherwise the solution also produced an unacceptable portfolio consisting mostly of cash and bonds. I also had to constrain the allocation percentages to positive values otherwise the solution occasionally included negative allocations indicating selling the asset short rather than buying.
Of course future performance rarely measures up fully to past results. While historical relations between asset classes may provide a reasonable guide, rates of return are often less predictable. In addition, as you can see from Fig. One solution is to rebalance the portfolio on a set time period to take into account the most recent correlations in order to maintain the desired level of risk exposure.
This method of asset allocation, is not the only one, however. A different, dynamic rather than static, approach would involve changing asset weights depending on market conditions. This can be accomplished by reducing the allotment of equities in favor of cash, precious metals or foreign exchange in a down market. A dynamic asset allocation trading system is also discussed in Chapter 16 of this book.
Values in between indicate the degree of correlation and their interpretation is subjective depending to some extent on the variables under consideration. The interpretation is different for medical research, social, economic or financial time series data. In the case of financial time series the interpretation can again be different depending on whether we compare raw price data or percent changes yields , as the direct calculation of the correlation based on absolute prices tends to overestimate the correlation coefficient as relations between financial price series are seldom linear.
Correlations based on price percent changes, on the other hand, produce more realistic values for the correlation coefficient as they deviate less from linearity. Therefore, although the correlation coefficient between two time series is unique, two different interpretations are included in Table 2. Table 2. The second column applies to the correlation between raw price data and the last column to percent weekly changes or yields.
Correlation coefficient r Interpretation Absolute value Price comparison Percent changes 0. If the correlation coefficient is squared, the result, commonly known as r2 or r square or coefficient of determination see also Section 2. In thus squaring correlations and transforming the result to percentage terms we are in a better position to evaluate a particular correlation.
There is also another factor we must consider when we try to interpret the correlation coefficient — the number of points we have used. If we plotted only two points, then we would be bound to get a straight line between them. With three points there is still a chance that the points will lie close to a straight line, purely by chance. Clearly, a high correlation coefficient on only a few points is not very meaningful. Traders often need to know if time series of commodity or stock prices are cyclic and, if they are, the extent of the cycle.
The correlation coefficient can also be used in this case by testing for auto-correlation at different lags testing whether values in a given series are related to other values in the same series. By doing many correlations with differing lags, the extent or duration of the cycle can be determined. It represents the percent of the data that is the closest to the line of best fit. In regression, the coefficient of determination is useful in determining how well the regression line approximates the real data points but it can also be used as explained above to interpret the correlation coefficient.
When this assumption is not true, the calculated value can be misleading. In practice this assumption can virtually never be confirmed; fortunately, the correlation coefficient is not greatly affected by minor deviations from linearity. However, it is always prudent to look at a scatterplot of the variables of interest before making important conclusions regarding the relation between two variables.
To understand what a linear relationship is, consider the scatterplots in Figs. The first scatterplot in Fig. Keep in mind that in the financial markets there is no such thing as a perfect linear relationship. In contrast, the plot in Fig. The variables X and Y are the daily percent change in gold and the dollar index respectively from 8 December to 29 December The relationship is approximately linear, as the points fall generally along a straight line. A chart of the two indices in Fig. This is the reason for the breakout of the linear relationship as depicted in Fig. The third scatterplot in Fig.
The period from to under study can be divided into six types. The relationship is not linear, as the best fit line is curvilinear. The first five from to marked A, B, C, D and E feature alternate positive and negative correlations and the last one marked F in Fig. The positive correlation periods can be allocated to area A in Fig. Area C consists of smaller periods with zero or near zero correlation enclosed inside the longer periods.
Eta is then calculated by dividing the group by the total variance. Eta can be used to measure the degree of linearity of a relationship, as the extent to which Eta is greater than r is an estimate of the extent to which the data relationship is nonlinear. Unfortunately, Eta is not particularly useful for analyzing financial markets because, in the case of time series, the groups should be categorized according to time and not rank.
Correlation 25 The relationship is not linear, as the points deviate significantly from the best fit line. Fortunately there are methods more appropriate for analyzing the nonlinear nature of the financial markets: artificial neural networks and kernel regression. Neural networks are nonparametric models which are inherently nonlinear in their computational elements and provide potentially the greatest opportunity for exploiting time series with low correlation dimension estimates by analyzing complex nonlinear relationships without making prior assumptions about the data distribution.
More information on neural networks is provided in Chapters 10 and In the case of most financial series this assumption is not justified, and a nonparametric measure such as the Spearman rank correlation coefficient might be more appropriate. Overlaying the normal Gaussian distribution we notice that these are not fully described by the Gaussian model.
The first is that the vast majority of the returns tend to be located near the center of the distribution and the second — and most important — is that the actual distribution has fatter and longer tails. This means that the probability of a larger movement in the index is much higher than that predicted by the normal distribution. The Gaussian model predicted a virtually nil 0. This is more obvious in the normal plot in Fig. The effect of the abolition of the uptick rule can be seen Correlation 27 Figure 2.
In addition, the standard deviation of daily changes, which was only 1. It is not certain, however, that the increased volatility can be wholly contributed to the repeal of the uptick rule as it coincided with the sub prime mortgage financial upheaval and the collapse of Bear Stearns. In any case, if we have to use statistical metrics that assume normality, it is preferable to use daily or weekly changes instead of raw index values.
In fact, Fig. Notice the unusually high frequency of more than —2. A single outlier is capable of changing the slope of the regression line considerably and, consequently, the value of the correlation, as demonstrated in the following example in Fig. By removing just one outlier on the far right of the graph , the correlation improved from 0. Nymex oil futures. By removing just one outlier enclosed in the square on the far right of the graph , the correlation improved from 0.
Correlation 2. Otherwise the correlation coefficient is a misleading average of points of higher and lower correlation.
A set of random variables having the same variance is called homoscedastic. Serious violations in homoscedasticity assuming a distribution of data is homoscedastic when in actuality it is heteroscedastic result in underemphasizing the Pearson coefficient. Heteroscedasticity does not invalidate the analysis and can be usually rectified by transforming the price data to yields or logarithms.
An example of heteroscedasticity is depicted in Fig. The plot shows a violation of the assumption of homoscedasticity. For lower values on the TSX-axis, the points are all very near the regression line, while for higher values the variability around the regression line increases dramatically. These could be caused by a number of economic or geopolitical reasons which affect both indices. Estimates for the values of a and b can be derived by the method of ordinary least squares. Table 3.
Extending this concept, the geometry of multiple regression where two or more predictor variables are involved , would involve fitting a plane in multi-dimensional space. However, since we live in a three-dimensional world, we cannot visualize the geometry when more than two independent variables are involved. We can extend, however, the same method only mathematically.
The practical problem in multiple linear regression is to select an effective set of predictor variables which will maximize the coefficient of determination, r squared, which is the proportion of variance in the dependent variable that can be explained by the variance of the predictor variables. Therefore, we want to include predictor variables that are highly correlated with the dependent variable but have low correlations among themselves. Row 21 22 23 24 0. The formula for calculating the other statistical metrics can be found in Table 2. The variables X and Y are the one day percentage change of gold spot and the dollar index respectively from 8 December to 29 December The more points fall inside the regression plane the better the predictive power of the regression model.
As a rule of thumb, intercorrelation among the independents above 0. The statistically preferred method of assessing multicollinearity is to calculate the tolerance coefficient for each independent variable. There will be as many tolerance coefficients as there are independents. The higher the intercorrelation of the independents, the more the tolerance will approach zero.
As a rule of thumb, if tolerance is less than 0. High multicollinearity can be ignored when two or more independents are components of an index, and high intercorrelation among them is intentional and desirable. These are usually combined in an index prior to running regression, but when this is not practically possible they can be entered individually.
Simple Eur/jpy Bitcoin Profit Trading Strategy -
For example, if data for an obscure index are not readily available some of the component stocks can be used instead of the index itself. Regression 37 Table 3. Because of normality and linearity problems associated with raw index prices, these are converted to weekly percentage yields and the correlation matrix is calculated depicted in Table 3.
The next step is to add to the regression equation the predictor variables one at a time and calculate the coefficient of determination r squared and the tolerance for each variable. As you can see in Table 3. The first model includes only the Euro Stoxx, the second includes the Euro Stoxx and the VIX and one more index is added for each subsequent model. The most important assumption is that of linearity. Checking that the linearity assumption is met is an essential task before using a regression model, as substantial violation of linearity means regression results may be more or less unusable.
Simple inspection of scatterplots is a common, if non-statistical, method of determining if nonlinearity exists in a relationship. An alternative is to fit a preliminary linear regression and to use the appropriate diagnostic plots to detect departures from linearity. Transforming one of the variables for example by taking differences or logs can sometimes help linearize a nonlinear relationship between two financial series. Normality is also a problem when considering raw stock or index prices but it is less of a problem when taking log differences or percentage yields.
The problem of the longer tails discussed in Chapter 2 can be partially overcome by removing some of the most extreme outliers. Normality can be visually assessed by looking at a histogram of frequencies. Regression 39 A more accurate estimate of the prediction is usually obtained by using nonparametric regression in cases of severe violations of linearity at the expense of much greater computation and a more difficult-to-understand result. Two common methods of nonparametric regression are kernel regression and smoothing splines. The smoothing splines method minimizes the sum of squared residuals, adding a term which penalizes the roughness of the fit.
SVM-based techniques typically use a kernel function to find an optimal separating hyperplane so as to separate two classes of patterns with maximal margin.
- - Intermarket Trading Strategies by Markos Katsanos.
- Handbook of Pharmacology and Physiology in Anesthetic Practice.
- Ethics and the Rule of Law.
- Shop now and earn 2 points per $1!
- Intermarket Trading Strategies!
- Invasion Ecology?
SVM models are closely related to neural network models. In recent years, SVMs have received considerable attention because of their superior performance in forecasting high noise and non-stationary financial time series. A major advantage of SVMs over other nonparametric methods is that they can have good performance even in problems with a large number of inputs. However, unlike other nonparametric methods the major difficulty of the SVM approach lies in the selection of its kernel, as choosing different kernel functions will produce different SVMs.
The criteria for weighting the stocks in the index are trading volume and market capitalization. The DAX is a performance index, which means that all income from dividends and bonus distributions is reinvested into the index. This is equivalent to the US Russell index. In simpler terms the volatility indices measure in percentage terms whether the DAX options are selling above or below their fair values as generally estimated by the Black-Scholes formula. The DAX is a highly cyclical index that now May faces a potential downturn as fully one-fifth of the index is made up of financial stocks, which are sinking to multi-year lows.
Tracking a sample of blue chip stocks, its performance is closely correlated to that of the market as a whole. The index contains 40 stocks selected from among the top by market capitalization and the most active stocks listed on Euronext Paris, and it is the underlying asset for options and futures contracts. The base value was on 31 December and historical data are not available prior to that date. Since 1 December , this index is no longer weighted by the total market capitalization of the component stocks but by their free float adjusted market capitalization.
This method of calculation, already used for other major indices around the world, ensures greater coherence between the real allocation of companies on the market and how it is expressed in the indices. It also limits the manifestation of volatility caused by too much of an imbalance between the weight of a stock in the index and the corresponding free float or available shares in the market. MONEP handles equity options, long- and short-term options, and index futures. Futures are liquid for trading purposes. An average of contracts per day changed hands for the earliest expiring contract.
The index calculation began on 3 January and historical data are not available prior to that date. Component companies must meet a number of requirements set out by the FTSE Group, including having a full listing on the London Stock Exchange and meeting certain tests on free float and liquidity. The constituents of the FTSE are reviewed and changed four times a year. Trading lasts from — when the closing auction starts , and closing values are taken at The highest value of the index to date was However, financials at The FTSE index reached a 5-year high on 13 July ; this was partly caused by the increase in the price of oil and the resultant increase in the share prices of BP and Shell.
The index is seen as a barometer of success of the British economy and is a leading share index in Europe. The criteria for including a company in the index are market capitalization and trading volume of the European companies. Both indices are market capitalizationweighted. The Dow Jones Euro Stoxx 50 Price was developed with a base value of as of 31 December and uses float shares. It is a priceweighted average the unit is yen , and the components are reviewed once a year.
The Nikkei Stock Average is the average price of stocks traded on the first section of the Tokyo Stock Exchange, but it is different from a simple average in that the divisor is adjusted to maintain continuity and reduce the effect of external factors not directly related to the market. At the height of the Japanese market in , Japanese banks were lending money secured by real estate. Investors took the borrowed money and ploughed it into the stock market.
A snowball effect began when the real estate market crashed, pressuring further property and stock prices. The Nikkei finally made a bottom on 28 April at On 12 August it broke out from a month rectangle formation and on 26 February rose to a six-year high of 18 The Osaka exchange has recently introduced the Nikkei mini contract symbol: NM. You can see that trading in some Asian markets is not very convenient for European traders as they open too late at night and close too early in the morning.
I thought it would be useful to include a list of symbols used by the most popular data providers or brokers for some international indices in Table 4. All the times shown are standard winter times. Please note that Japan and China do not currently observe the daylight saving time DST or summer time and consequently the CET and EST opening and closing time indicated should be moved forward by one hour. The transition to summer time starts on 25 March and ends on 28 October in Europe.
In the United States the transition starts two weeks earlier, on 11 March and ends on 4 November. You will notice that US brokers are still of the old fashioned mentality that foreign markets are irrelevant to US investors and do not provide any data for non-US indices. Of course US markets do not exist in a vacuum and it would be very helpful for US investors to check how the international markets are doing before the US market opens for trading.
I thought it would be useful for international traders to include the most likely daily percentage changes for some of the major international indices and commodities in order to assess risk adjusted positions in each market accordingly see Table 4. Each value in Table 4. Frequency distribution values are included instead of standard deviations because, as discussed in Chapter 2, none of the indices conform to the normality assumption. An interesting comparison between actual observations and theoretical probabilities as predicted by the normal distribution is depicted in Table 4.
It can be seen that there is an exponential shift from the middle to the tails of the distribution making predictions beyond the 0. This 48 Intermarket Trading Strategies Table 4. They usually differ from provider to provider by a prefix or a suffix. Blank cells indicate that the index was not available at that broker or data provider.
The daily changes are divided by percentiles according to values below which certain percentages of cases fall. This rule has been in effect for 78 years. But can the recent volatility in the markets be blamed entirely on the rescinding of the uptick rule? At the time of writing, not enough data were available for such a study as the recent volatility has been exaggerated by the combination of the housing and related financial upheavals.
The last column is the day average volume as of April The above formula 4. An example of calculating the dollar index on 8 February using formula 4. The FX rates from Bloomberg are raised to the negative of their weight, and the product, when multiplied by the index constant of The dollar index component currencies and their weightings are: Euro These are depicted in the pie chart in Fig. Intraday values for the dollar index are generally not available from data providers.
From the end of World War II until the early s, the dollar was tied to other major currencies by the Bretton Woods fixed exchange rate agreement. After it became freefloating, the dollar went through a roller-coaster ride. It rose to an all-time high of in but then retreated to almost where it started. At the beginning of it rose again to , making a year high and more recently, in , mainly because of the huge US trade deficit, it plummeted to an all-time low of While data on tradeable instruments such as bond yields, the dollar, international bonds etc.
Table 4. I included both a short and a long time span because, as it becomes apparent from Table 4. This was more pronounced in the case of the CRB index, crude oil and gold, mainly because investors turned to commodities as a hedge against the recent disorderly dollar decline. The British pound, the euro and the yen had the highest negative correlation in both time spans, but this was expected as they are all components of the dollar index.
Of course this is nothing new and it is normal for gold, which is denominated in dollars, to have a negative correlation with the dollar. Surprisingly, the dollar index correlated better with the Australian dollar, which is not a dollar index component, than with Gold. Lastly, the dollar index did not correlate well with equity indices. The equity index that correlated best with the dollar index was International Indices and Commodities 53 Table 4.