There is one factor which sums up the decision of the smart money traders to throw in their money in one direction, or to buy into a currency over and above another currency: sentiment. Sentiment means developing a bias for something or against something. When market sentiment towards a currency has shifted either in its favour or against it, you better pay attention and fall in line.
This article is all about sentiment detection, and how it can be used for predicting the direction of flow of a currency. Sentiment analysis in forex is what fundamental analysis is all about.
Definition of Market Sentiment
A key component of success in forex trading is the need to be able to anticipate and therefore predict the direction that a currency pair will go, and in good time too. To be able to make such a decision, traders must factor in the sentiment prevailing in the market.
But what really do we mean by market sentiment? In order to keep it simple, we will simply refer to market sentiment as the opinion of the markets by the traders who control the largest volumes of trade. Notice we did not say the opinion of a large number of traders, because the opinions of the majority of traders in the market are generally not correct. However, the smart money players who are in the minority numerically but who control most of the money (remember Pareto’s 80:20 principle) in the market are the ones whose opinions count. Their opinions will constitute the market sentiment.
So how can traders detect it? For traders to be able to detect it, there has to be a shift from the old ways of viewing the market and get down to using the new tools available to outline the true sentiment that will move currency pairs. The new reality is that trading activity in the 21st century is now a social web phenomenon. The Internet has had a profound impact on trading due to its ability to propagate information, and in a very fast and efficient way which no other medium can. As such, it is easy for a large mass of traders to get drawn into an opinion. According to a study done by Adamatzky in 2005, the fusion of individual minds into one collective mind constitutes a crowd mind. When this crowd mind forms, then the individuals in the crowd lose their individuality and start to behave like a herd, with mostly irrational and impulsive behavior. This sums up the behavior which is seen when lots of traders rush into positions like a herd, even when the rational traders have seen the danger signs and are heading out of the market.
The emergence of the Internet is producing new sets of tools that can be put to effective use by the forex trader. One of these modern tools is known as text mining. Text mining is the practice of finding word trends in a given document with a goal of determining the meaning of those documents and text.
Text mining can be used for detecting market sentiment. Text mining uses the search engines, instantly scanning sites for useful information about what the real experts are saying about a currency pair way before such events take place and converting that information into a useful trading nugget. I recall sometime in 2008 when a famous banker in a large Japanese bank had predicted that the Euro, which was approaching 1.6000 to the USD at the time, would actually fall to 1.4000. It looked so far-fetched at the time but look where we are today: The EURUSD had even gone as low as 1.1900 at the height of the Eurozone sovereign debt crises. Mining the internet for such information in the context of text mining could form the basis of detecting what the prevailing sentiment is among the smart money players at any point in time.
Closely related to text mining is the use of social media tools such as Twitter and Facebook to profile mood changes in the forex market. A statistically reliable and effective Twitter data processing tool is found on twittersentiment.appspot.com. This site is able to take a snapshot of Twitter opinion so the trader can see the data stats in a usable form i.e. in a form which can enable the trader to gauge market opinion.
Once market sentiment is obtained, the next logical step would be to match it to the price action of the currency pair that the trader is interested in trading.
In another example of matching sentiment against price action we can see how positive and negative sentiment regarding crude oil correlates with price action (Figure 3.2). News, blogs, video, and forums over the Internet were scanned during the week of May 28 to June 9 on sentiment regarding crude oil prices. Each day’s negative scores were converted into a line graph and overlaid against each day’s positive scores. Then the actual Brent crude oil prices were matched against these negative and positive lines. While this is only a sample period, we can see that it is worthwhile to use sentiment data as a gauge for price direction. A peak in negative sentiment on crude oil occurred on May 31 as crude oil prices reached a high of 102.98. It was followed by decline in crude oil prices. Sentiment reached a bottom negative score on June 5 and positive sentiment started bouncing up. Crude oil prices rose back to the 102 area a few days later. While the data needs much more granularity, we can sense, even at this early stage in the art and science of sentiment-based signals, that there are two key areas that will be useful to the trader. First, when positive or negative sentiment crossover, the trader can use this as a clue that market opinion is shifting. Additionally, it appears that sentiment peaks, whether positive or negative, are the key milestones relating to subsequent price changes and offer great potential as a source of trading signals.
How to Apply Sentiment Detection Tools to Forex Trading
Let’s explore each key step on how any trader can apply sentiment detection tools.
Step 1: What is at Play in the Market? Risk Appetite or Risk Aversion?
The first step in applying sentiment detection tools to forex trading is to decide on what market sentiment is to be monitored. What emotion is currently at work in the market? At any point in time, one of two major emotional forces will be at work in the market:
- risk appetite
- risk aversion.
Now it does not mean that one is good and the other is bad, because in forex, you can make money from good news AND bad news. It just depends on you the trader being on the right side of the coin. At any point in time, traders will either have a risk appetite (in capital enhancement mode), or they will be risk averse (in capital preservation mode).
In 21st century trading, sentiment outweighs economics in impacting currency price movements. This is not a failure of fundamentals in the equation, but it means that the market also works on expectations. The words “risk appetite” connotes market optimism, while risk aversion connotes fear. Every day, there is a competition between market optimism and fear, and the constant shifts in the delicate balance to one side or the other will dictate which of them will take hold. So any time the forex trader wants to take a position, he or she is in effect, measuring the emotions at play in the market.
Market direction reflects a precarious balance of optimism and fears. There are many market fears and any one of these fears can take hold of the market very quickly. Not all fears are bad for a currency. For instance, inflationary fears will generally lead to interest rates going up, which will fuel bullishness on commodity prices and the commodity currencies, increasing market optimism or risk appetite. Fear of turmoil or war in the Middle East causes crude prices to go up, as well as the price of the commodity currencies backed by crude such as the Canadian Dollar. On the other hand, fear of a drop in the GDP of China will lead to bearishness on the commodity-backed currencies such as the Australian dollar, and risk aversion will then take hold of the market.
So it is important for traders to be familiar with the various fears at work in the market. So how can a trader scan the markets to see which fears are operational at any point in time?
Step 2: Scan for Specific Fears
Before the trading week commences, the forex trader must scan the market and determine which of the market fears is dominant. Correctly answering this question will lead correct identification of the market direction for the week. This is where action analysis comes in, using the internet to search for information. According to Oberlechner, financial news reports usually consist of the perceptions and market interpretations of the trading participants, which are fed back to the traders in the market. The forex trader’s job is to filter out the wheat from the chaff. It is not rocket science.
Consider the recent upgrade of Spain by credit rating firm Fitch. Spain had been downgraded in 2012 and this caused the Euro to take a massive hit. As a currency without much market fundamentals except those which confirm the strength and stability of the Eurozone and its financial system, such a news event as a credit rating upgrade is one which would definitely bring on risk appetite for investing in the Eurozone and its currency. Of all the market fundamentals at play, a discerning trader can zero down on this one and make rational trading decisions on the Euro. The next step would be to pick out the best currency pairing of the Euro to trade this news release. This would lead the trader to again mine through all the text out there to see which counter currency in a Euro pairing has a negative sentiment. A long position on the positively impacted Euro against a negatively impacted counter currency would be the play here.
Step 3: Scan Headlines
Could the average forex trader use sentiment mining to help shape their trades? If so, what everyday tools can be used by the trader to accurately extract market sentiment? The challenge is to spot occurrences of key terms that are tagged to the fundamental forces being searched. The main idea is to find the right terms. Many traders overlook the value of scanning headlines. Headlines provide an avenue to capture fundamental opinion. We just talked about the credit upgrade on Spain by Fitch. This is an example of a screaming headline which a trader must scan for. Headlines are effective because they are specifically designed to catch attention. They may not be very accurate in their representation of the actual economic data, but they are effective in showing the pulse of opinion. Headlines trigger excitement in contagious fashion. Headlines can amplify sentiment.
An example of this was seen Standard & Poor’s downgraded the credit rating of the U.S. government, triggering a huge market response. A keyword used in the headline was the “Negative”. Even though this may have exaggerated the situation, it did do the job of drawing the market response and whipping up a market sentiment.
Step 4: Form Your Own Keywords for Search Retrieval
Another important sentiment detection tool in the hands of the forex trader is keywords. Knowing what keywords to use when conducting text mining on the internet is a good way of measuring the emotional strength of any optimism or fear operating in the market on a continuous basis. The use of keywords allows traders find the character of the market sentiment.
Keywords used must be specific. They must not be too generic as to include unnecessary stuff that will add confusion to the mix. Terms such as risk appetite, risk aversion, and inflation are good terms because they sum up factors that are always at play in whipping up market sentiments. Keywords must also be able to link to emotions and categorize them as positive and negative emotions.
Let’s apply this to the forex trader scanning the web using keyword groups. The trader can start by linking the names of the central bank chiefs to emotions. For example, a trader may use the keywords “Draghi admits..” A context in which this could be used was when the free-falling Euro needed a real boost, and after months of equivocation, ECB Chairman Mario Draghi finally admitted that the ECB needed to buy Eurobonds to help stabilize the Euro. This gave the Euro a much needed uplift. Now supposing a trader used the keywords as stated above. The chances are that the news event detailing Draghi’s statement would come up on the internet search and the trader would immediately know what market sentiment would be.
To retrieve the latest emotional content on inflation and the New Zealand Dollar, the trader can use a different group of words. Inflation fears or the Kiwi Dollar could produce a quick grab of words that relate to positive emotions about the NZD. The result is a greater detection of emotions. As a general rule, combine the name of a key market leader or the name of a central bank with words such as: admits, declares, warns, supports, denies, etc. The result is a retrieval of news that carries with it a lot of information about the emotions involved. A good idea for any trader would be to create their own table of their own opinion seed words, or a verbal quadrant. For instance, a look at the most recent Rate Statement by the Reserve Bank of New Zealand would indicate the association the bank made to rising immigration, increased housing demand, increase in consumer prices and inflationary fears, leading the RBNZ to hike interest rates from 2.75% to 3%, thus further increasing the profitability of carry trades between the NZD and USD or JPY. This would immediately put in play positive emotions on the NZD pairings.
The effect is to provide a better match between the search and the retrieval of the emotion involved. Once an underlying currency pair is chosen, word searches should become more targeted and specific. Here are suggested words for use at any time for trying to gauge market sentiment.
Start with any underlying market and add the suggested key words:
- using the following world combination formula:
- underlying market + risk appetite
- underlying market + risk aversion
- underlying market + fears
- underlying market + optimism
- underlying market + doubts
- underlying market + pessimism
We can generalize the entire search process in one equation or algorithm. It would be: underlying market + emotional adjective or adverb. To be clear, a trader would enter at different times:
- U.S. Dollar Index risk appetite;
- U.S. Dollar Index risk aversion;
- U.S. Dollar Index fears;
- U.S. Dollar Index doubts;
- U.S. Dollar Index optimism.
The result is an ability to count the positive and negative news items relating to the U.S. Dollar Index. This follows recent text mining and sentiment analysis methodology.
Step 5: Create Your Own Risk Appetite/Risk Aversion Ratio
Since the words risk appetite and risk aversion are extremely effective as seed words to retrieve market sentiment, the trader should always conduct a general risk appetite and risk aversion search. By comparing the results of the search, the ratio between risk appetite and risk aversion beliefs can be approximated.
After retrieving the results of using risk appetite and risk aversion, the next step is to create your own sentiment ratio. The result is your own ability to detect if the mood of the market is up or down. The risk appetite /risk aversion ratio compares positive to negative sentiment results from your own text searches.
After doing a search a good way to quantify the balance of risk appetite to risk aversion is to assign a number to the article or headline. Ask yourself: Is the article retrieved positive or negative about the underlying market? Keep score. This helps keep track of the strength of the sentiment. A useful technique is to use a ranking range of –5 to +5. If the total sum is positive, in effect, you have a market that is risk positive.