Quantitative Analysis in Forex
Quantitative analysis allows traders to remove emotion from the investing process. Quantitative analysis is an approach that focuses on statistics or probabilities over gut feelings. Given the technology of computers and sophisticated math models, quantitative analysis has taken over Wall Street and a majority of new traders and employees at Wall Streets or those with a quantitative mindset. Quantitative analysis has a place in the FX market just like any other market.
You are likely familiar with different forms of quantitative analysis even if you do not consider yourself a quant, which is someone that approaches markets from a quantitative standpoint. A simple financial ratio such as earnings-per-share or something more difficult like options pricing and discounted cash flow are forms of quantitative analysis. As you can imagine, data is critical in the analysis is often only as good as the data going in so many quants focus on the quality of data used to fill out their mathematical and statistical models.
Examples of Quantitative or Statistical Analysis
You don’t have to be a math whiz or have a doctorate in econometrics to benefit from statistical analysis. With statistics, you are looking at dependence or association of two random variables or to datasets. Traders benefit from the common statistical analysis of correlations, which refer to a broad class of statistical relationships and dependence. A common correlation in the FX market is dollar weakness is correlated with a weakness to emerging markets. Another intermarket relationship yen strength and equity market weakness.
Statistical analysis is helpful in determining future probabilities but is not meant to be purely predictive. A typical statement is that correlation is not causality. Causality means explicit cause-and-effect, whereas correlation simply means potential common movements between two random variables. The scale of correlations coefficients is -1 to +1 whereas the negative one is a perfect inverse relationship or correlation, zero is zero correlation, and a positive one is perfect positive correlation almost like the two variables or markets are handcuffed to each other.
Another favorable form of statistical analysis is known as regression analysis. Regression analysis is a very favorable statistical model and quantitative analysis in order to help you see the relationship among variables. Regression analysis focuses on the relationship between a dependent variable and one or more dependent variables. Specifically, regression analysis helps you to understand how the typical value of the dependent variable changes when any one of the independent variables as varied. Most FX charting packages have a regression channel that does the calculation of regression analysis for you and is often easier to access than correlations.
Regression analysis commonly estimates the conditional expectation or direction of the price of the dependent variable given the independent variable. This means the average value of the dependent variable relative to a fixed independent variable. This is often shown in a sloping line higher or lower cutting through price in the direction of the trend or in a sideways move the regression line is often flat.
What Is Needed?
While mathematical models are beyond the scope of this article, many traders utilize Excel from Microsoft and use the correlation function between the variables over a particular set of time to determine if there is a positive or negative correlation. However, many research outlets will put out correlation reports and they can also be found on research terminals like Bloomberg.
If you are interested in doing these types of models yourself, it’s important to note the results are data driven and missing or incomplete data may lead you astray. Therefore, you should take care of the missing data first in order to have an effective analysis of the data. Excel is likely your best bet in terms of doing the simple analysis but many brokers provide tools that can help you do a lot of the analysis as well.
In conclusion, statistical analysis is meant to wrap your head around seemingly random variables for a pattern that you can trade. Risk must always be managed, but these patterns can last for a long time even without causality existing. While seemingly similar, backtesting is the proverbial wolf in sheep’s clothing of often statistical or quantitative analysis. It pays to be aware of backtesting pitched as statistical modeling because more often than not backtesting is done using over-idealized data sets which can bring about false confidence, over-leveraging, and potentially large losses when the current environment diverges from the data set.