Week 31: Trading Events

Fun and scary and profitable and dangerous

Oct 12, 2022

Hi. Welcome back to Fifty Trades in Fifty Weeks!

This is Week 31

50in50 uses the case study method to go through one real-time trade in detail, about once per week. This Substack is targeted at traders with 0 to 5 years of experience, but I hope that pros will find it valuable too. For a full description of what this is (and who I am), see here.

If you want to learn about global macro in real-time … subscribe to my daily: am/FX

Listen to this as a podcast on the web … or Spotify … or Apple.

Update on previous trades

I was in Brazil (São Paulo and Rio) for a bit, but I’m back!

The USDJPY trade I outlined for 50in50, Week 30 was either great or terrible, depending on how you played it. I stupidly forgot to put the take profit in the write-up. So, I’m marking it as a loss because even though USDJPY rallied significantly after I went long, it got smoked by MOF intervention. And since I forgot to communicate my take profit: I take the L.

Sometimes in trading, little screw-ups like forgetting to cancel an order, or leaving an order as DAY instead of GTC by accident can cost you money. Outtrades and screw-ups are a cost of doing business. Minimize them, and cut them as soon as you discover them. And don’t be too hard on yourself.

One time I left an automated stop loss on a machine for 70 million euros in the morning, to protect a long position. The day went on and about 5 hours later, I squared up my position for a decent profit.

I went for a run outside, and when I got back, EURUSD was unchanged but my P&L was $500,000 worse than when I left and I was short 70 million euros 70 pips out of the money. Turned out there had been some clickbait bearish headline about a downgrade of Spain (it was during the eurozone crisis) and EURUSD went down 100 and up 100 while I was off desk, stopping me into an unwanted short euro position and pushing me 70 pips underwater.

Horrible.

Anyway, outtrades and errors happen. I have an even worse story than that one! Forgive yourself and modify your process so you don’t make the same mistake twice.

Now let’s talk about event trading.

Note: I am sending this right at 4 pm NY, the market close, on the day before CPI. That will mean most people will read it after it’s too late to do the exact trade I described. That’s OK! This is supposed to be a learning exercise, not a trade copying exercise.

Trading Events

Once you have a solid understanding of the fundamentals that drive your market and you have a good handle on the current narratives, you are ready to trade events. With your knowledge of current themes along with positioning and sentiment, you are ready to figure out how new information will move prices. Remember that the current price is the sum total of all the known information in the world (public and private).

When news comes out (e.g., economic data, monetary policy announcements, or unscheduled headlines) the price zigs and zags as it looks for a new equilibrium level. This search for equilibrium is sometimes said to be guided by an “invisible hand” because it is uncanny how price rapidly incorporates new information and adjusts to the new outlook.

When news hits, there is a significant opportunity for profit. Traders that have the best understanding of the fundamental backstory going into an event will have the best chance of coming out the other side in the black. Before you trade the news, you need to have a solid understanding of what is driving the markets you trade and any correlated markets.

The prevailing economic and fundamental landscape has a massive influence on how the market interprets new information. For example, during the 2010/2012 period, the stock market consistently rallied on bad economic news because it was perceived that bad economic news meant the Fed would announce more quantitative easing (QE) and more QE usually means higher stock prices.

Traders who weren’t fully cognizant of the macroeconomic landscape in those days were always complaining “This price action makes no sense! Why are we rallying on bad news?!?” Those that were focused and did their homework understood the reaction function perfectly. A nuanced understanding of the market gives you a huge edge when trading the news.

There are many types of news and events that can generate profitable trading opportunities. The main ones are:

Economic data
Central bank meetings
Corporate earnings
Central bank speeches and comments
Other headlines and news events (corporate announcements, mergers, geopolitical news, etc.)

CRITICAL CONCEPT: Understand What Is Priced In

Before you can trade a data release, central bank meeting, or any other event, you need to understand what is priced in. This is an absolutely critical concept for anyone who trades any financial market. If you trade or you are ever planning on trading any security in any market, you need to fully and completely grok the concept of “Priced In”. If you read this section and still don’t really get it, ask someone to explain it more fully. Do not trade events until you are completely comfortable with the concept of what is priced into a market.

Financial market prices are generally set by the sum total of all the buying and selling done by every human and computer market participant. Different buyers and sellers will decide to buy or sell depending on the price, and as prices move, different actors with different time frames will enter or exit the market.

Decisions to buy and sell are generally driven by the market’s assessment of fair value for the security and this assessment should (in theory) be composed of all the available information that relates to that security. As new information comes in, prices move in real-time and should reflect and anticipate all available information, often before the information itself is released. Let’s look at an example:

Currencies are highly sensitive to interest rate movements. Therefore, it makes sense that currencies should move when central banks adjust their primary interest rate. Here is an example of how a currency can move when a central bank unexpectedly changes interest rates.

Hourly USDCAD around a surprise interest rate cut from the Bank of Canada (January 2015)

Note: USDCAD higher means weaker CAD and stronger USD

When something unexpected like this happens, it is not priced in. This move higher is what you would anticipate because an unexpected rate cut makes holding Canadian dollars less attractive given the lower yield. The market sells CAD and buys USD, driving USDCAD higher.

This was NOT PRICED IN. In contrast, let’s look at an interest rate move that was priced in.

In 2008, The Reserve Bank of New Zealand (RBNZ) cut rates dramatically in response to the global financial crisis and then they cut rates again after the 2011 Christchurch earthquake. By 2013, economic activity had rebounded significantly and it was clear to the RBNZ that rates were too low. Inflation was picking up and the RBNZ felt it would become appropriate to start raising rates at some juncture, but not right away. The story plays out as shown in the chart below. The first hint that the RBNZ wanted to hike rates was delivered on July 25, 2013, and it is marked by point 1 on the chart.

The RBNZ kept warning more and more loudly that rate hikes were coming, so NZD strengthened vs. the AUD (AUDNZD went down).

Daily AUDNZD, June 2013 to July 2014

AUDNZD down means weaker AUD, stronger NZD

The pair falls as the market anticipates a NZ rate hike then rallies when the RBNZ hikes

“Although removal of monetary stimulus will likely be needed in the future, we expect to keep the OCR unchanged through the end of the year.” – July 25, 2013
“OCR increases will likely be required next year.”—September 2013
“Although we expect to keep the OCR unchanged in 2013, OCR increases will likely be required next year.” – October 2013
“The Bank will increase the OCR as needed in order to keep future average inflation near the 2 percent target midpoint.” – December 2013
RBNZ hiked on March 13, 2014

This concept of the market pricing in an event and then reversing (somewhat counterintuitively) after the event is known as “buy the rumor / sell the fact”. We covered that concept in Week 28 as a reason to go short ETH at 1650. That idea worked out well.

You can see that the RBNZ slowly prepared the market for the interest rate hike over a period of 8 months and the NZD gradually appreciated over this time period (lower AUD, higher NZD). Finally, the RBNZ hiked in March 2014 but it was completely expected—fully priced in. The market priced in the hike over the course of many months and when the hike finally arrived, NZD sold off in response.

This reaction is counterintuitive to someone who is not following the NZD closely. “They hiked and the currency sold off? What the???” But any professional following the NZD would know that the hike was fully priced in and so the reaction in the currency market is perfectly understandable. There is no need for NZD to rally after a rate hike that is fully priced in.

In fact, quite often the reaction to an event that was priced in is the opposite of what the macroeconomic textbooks might suggest because those that had been buying the NZD in anticipation of the rate hike sell (take profit) on the day of the event.

In every market event, the outcome is determined, to a significant degree, not by the event itself but by the combination of the outcome of the event and the market pricing. If something is 100% priced in, the market will not react when it happens. The interplay between expectations (what’s priced in) and reality is part of what makes trading events super complicated and interesting.

Now, a totally unbiased advertisement:

If you would like to receive am/FX in your inbox every day and get my global macro thoughts, live trade ideas, and some random silliness… Subscribe here.

This logo makes you want to subscribe to am/FX. This logo makes you want to subscribe to am/FX.

Trading Economic Data

In the “good old days” before algorithmic trading, you could make money trading economic data in the seconds right after the release. If a US data release was strong, you could just quickly buy 20, 30 or even 50 million USDJPY and then sell it out just as quickly to slower traders at other banks as it rallied. You could buy TY on a weak CPI and make money if you had quick reflexes honed by years of video gaming (I do).

Those days are long gone. There are now hundreds of algorithms programmed with code that reads something like:

If nonfarm payrolls headline number beats by 80,000 jobs or more, buy USDJPY at current market or up to 50 points higher
If NFP number beats by 150,000, buy USDJPY up to 100 points higher
If NFP number misses by X, sell Y… And so on.

Instead of gyrating back and forth for a while before finding equilibrium, the market now finds equilibrium super fast after the release of economic data. Hedge funds and banks use regression analysis before the data release to estimate how far each asset should move on various outcomes and they execute trades via algorithm at the moment of release. They will buy or sell until all asset classes are approximately in line with where history suggests they should be and this will happen in microseconds. There is nothing left on the table for the human traders because all the liquidity in the market is hoovered by the bots in the first few microseconds.

Imagine you are in a gymnasium with poor lighting, trying to shoot rats with a slingshot. There are a few other people in the room doing the same thing but you are a pretty good shot and your vision is good. You are doing OK. You are hitting your fair share of rats.

Suddenly, a cyborg marches into the gym. He is equipped with infrared night vision and a fully-automatic laser gun with motion-sensitive targeting. He shoots 50 rats in 10 seconds without missing. You and the other human rat hunters stand there, jaws on the floor.

That is what trading economic data is like now that algos are programmed to trade every economic release. With hundreds of electronic players all using similar algos, competition is at the microsecond level and humans are shut out. Pre-2005 or so, you could just about make your year trading economic data alone. Now there is far less money to be made by humans after data releases. This is one of many examples of how you need to adapt to changing market conditions over time. Adapt or die.

So now, instead of four strategies, there are three main strategies around economic data.

1. Go with. Trade the extreme data.

Algos will take out the first pocket of liquidity when economic data are released, but there are still opportunities for profit when important data come in extremely strong or extremely weak. This is because the bots can only buy or sell so much (they have risk limits too!). The first wave of algo trades can be followed by more waves of macro and hedge fund trades in the same direction and you can get in after the algos but before everyone else.

If you know what a big miss looks like, because you have studied all the past releases and printed out a little sheet showing 1, 2, 3, and 4-standard deviation misses for the data. And you know positioning and expectations and what is priced in (beyond the simple median economist forecast)… You can sometimes GO WITH extreme data points.

Last month’s CPI is a clear example of this. The market had killed all the shorts, bulls were getting feisty, and the talk was of peak inflation. The SPX had rallied from 3600 to 4300, right into the 200-day moving average. There was no room for error as momentum, positioning, and expectations were all one way. Then this happened:

S&P futures, 5-minute chart through CPI in July 2022

Oops! Any trader going into that CPI number with a good knowledge of pricing and a flat book could have made a ton of money simply selling S&P futures at 4075 or so after the number. It’s never an easy trade selling something down 100 handles, and this is only easy in hindsight. But it’s a textbook example of how you can go with the data in cases of extreme outliers into wrong-way positioning.

2. Beat the algos! Go the other way after economic data

Another way to profit from economic data is to take the other side (or fade) the initial kneejerk move. The term “fade” means to take the other side of a big move or view in the market. Sell a rally or buy a dip. Quite often on a trading floor you will hear someone say “This move is a fade” or “This move is baloney; I’m fading it.”

Fading economic data is a scary strategy because you are going against short-term momentum and against the theoretically logical direction dictated by the economic release. That said, with the proliferation of algos trading the initial headline, quite often the fade strategy is most profitable.

The reason fading the kneejerk move often works is that algos and humans pile in based on the initial reading but there are quite often times when the details are not as strong as the headline or the headline is not really all that strong in the first place. Then you often see a rapid turnaround in the market because the algos are risk-averse and so they hit or pay almost any price to get out of a trade that is not instantly in the money. As usual, it pays to know the setup going in. Which way is the market leaning? What is the asymmetry?

A key skill in trading is to identify asymmetries. Always be on the lookout for moments when the market looks much more likely to go a particular direction after an event due to positioning or some other bias in the market.

As a human trader, there is sometimes a profitable opportunity to take the other side of a move if the headline is not representative of the full story. For example, say the headline jobs number is extremely strong but the details of the report (unemployment rate, average hourly earnings) are weak. You can sell the spike in USDJPY because the bots will be long too many USD and will need to sell. The best way to be ready for this type of reaction is to listen to the immediate reaction and commentary from economists and strategists.

This is easiest if you work at a bank or hedge fund because you can hear your economist’s reaction almost immediately, but there is also real-time reaction on CNBC, Twitter, Reuters, Bloomberg, and so on. If the headline is very strong but the details are weak, the dollar will often fail to hold initial gains as the market views the knee-jerk as incorrect or exaggerated and fades it.

Remember that the last trade before the release, what I call the “NewsPivot”, is a simple but effective reference point for trading economic data and other news releases. Make a note of where each of the majors is trading before any major release and consider those levels to be the critical pivots for each pair for the rest of the day. This works because the reaction to the news is often just as important as the news itself.

If very good news comes out and the market goes back through the level it was trading before the news—that’s bearish price action. This is a concept known as good news/bad price, which we covered in Week 17. The opposite situation (bad news/good price) is bullish.

Why would a market rally on bad news? There are all sorts of reasons for counterintuitive moves after news and economic releases but the two main ones are 1) The underlying details of the release are not consistent with the headline release (strong headline/weak details) and 2) supply and demand, not news, determine market prices.

3. Take a view. Go into the event with a position.

I am confident from experience that the highest expected value way to play events is to go in flat and trade either strategy 1 or strategy 2 above. On the other hand, the highest returns come from betting before the number. Higher returns, but much higher risk and a lower hit ratio. Before you can take a position into an event, you need to estimate jump risk. That is, if you’re wrong… Where can you get out?

If you can’t estimate it with confidence, forget about the idea. Rule one of trading is: Eliminate risk of ruin. Don’t put on trades if you cannot quantify your risk. Don’t be short USDHKD. Don’t front run the underwriter’s stabilization bid after an IPO. Don’t go max short a stock through earnings, etc.

Estimating jump risk

There are two ways to estimate jump risk. The first is to look at historical data for a given event and study the maximum, average, and median moves. The Bloomberg function ECMI is useful for this type of analysis. Looking at how markets moved over the same event in the past is usually a good way to estimate gap risk but it is, by definition, backward-looking. A more accurate way to estimate jump risk is to use the options market.

The difference between volatility pricing for the day before an event and the day of an event can be used to extract an approximate jump risk for the currency. The calculation is beyond the scope of this piece. To be perfectly honest; when I want this information, I don’t calculate it myself, I just ask an options trader.

If you work at a bank or have an institutional relationship with a bank you can simply ask: “What is the options market pricing as jump risk for event X?” and they will give you an answer, usually in basis points. If not, you can ask someone smarter than me to get it for you.

Once you have a sense of the jump risk, you can take a cash position. If you are not comfortable with gap or jump risk, you can trade options. The issue there is that options through events tend to be overpriced because market makers charge a premium for the jump risk! You don’t want it and neither do they. If you want to buy it, you have to pay up. Still, if you’re right, options will often realize much more than the event premium anyway. And you can’t blow up.

Factors to consider

Going into an event, there are a few main factors to consider:

What is positioning? This will help you determine the weak side, or asymmetry through the number. If everyone is short, your market is more likely to see an impulsive topside move post-event.
What’s the FOMO trade? People hate to miss the “obvious” trade. For tomorrow’s CPI, the FOMO trade is strong CPI so people are positioning for that because they will feel stupid if they miss an epic inflation print.
What is the skew of the distribution for the event? Sometimes you can get a lean on the event using various methods. That’s beyond the scope of this piece, but a simple example would be to use a good model of the correlation between PPI misses and CPI misses to forecast the probability that CPI will miss one way or the other.
What is the most recent move? Generally, the move going into an event will tell you a lot about positioning. If your market rallied hard into nonfarm payrolls, it’s more likely traders are long that market, not short. Weak hands that want to square up, lotto ticket buyers, and those with a strong view on the data all wait til nearly the last minute before putting on their bets. As such, you get a read on the positioning and the lean by looking at how things moved beforehand.

The trade: Long FXA calls into CPI

The USD has been ripping all year, and the last CPI was a barnburner. Recency bias suggests that number is front of mind, and given the monster move after CPI last month, and the inflationary macro environment, fear of a strong CPI makes sense.

The issue is that the follow-through after a big runup like this one into CPI hasn’t been great this year. Obviously, the number itself matters tomorrow, but the conditions going in are important too.

Here's AUDUSD three days into, day of, and three days after CPI releases this year.

AUDUSD %return: 3 days before CPI, the day of CPI and 3 days after CPI

t-3 is total return of the three days before CPI. t is CPI day. t+3 is sum of the three days after.

Look at the dark blue bars. The USD ran up (lower AUDUSD) into most CPI releases. But then look at the lighter blue bars. The move often reversed on the day of CPI. The takeaway is that in months like January, March, April, and August, the big runup in the dollar was a fade once CPI came out. The only exception was June, where you had a 1% drop in Oz going in, and another 0.7% drop on the day of CPI.

The market has been extremely bullish USD most of the year and the USD has gone up most of the year, but price action around CPI and FOMC days has been much more dollar bearish. Dollar bulls have been adding to risk since nonfarm payrolls last week and I think tomorrow’s CPI could see a huge reversal. Also, there has been a firehose of bad news in GBP and GBPUSD can’t go down (bad news/good price). That could be a tell that the market is over its skis on long USD trades.

But I am too scared to get involved in spot because the risk of a strong number is still meaningful. When you trade events, there is never a sure thing. You are trying to find 50/50 bets that pay 2:1 or 60/40 bets that pay 1.8:1 and so on. It’s always probabilistic. Everything I do in trading is probabilistic.

When you have a probabilistic mindset, you know you are going to be wrong a lot and you get used to it. That makes it easy to cut your losses and easy to flip views and find new ideas. I have zero attachment to my trades. I never fall in love with them. I know some will work and some won’t. I just want to put myself in a ton of 60/40 bets that pay 1.5:1 and let the chips fall. Speaking of chips, I found a poker game when I was in Rio!

Check out this had. I had KK vs. AA and the flop was K 3 K. Lulz.

For tomorrow’s CPI release, I think the USD will sell off much more on a weak number, so while I have no counter-consensus forecast for CPI, the asymmetry makes me favor short USD.

AUDUSD is very often the best trade through US data because AUD is a high beta currency that trades consistently with risk appetite and general USD direction. You can trade AUDUSD via the ETF “FXA” and you can trade options on FXA. They are not super liquid, but they are liquid enough.

For this trade, I’m going to buy $2000 worth of 21OCT FXA 63 calls for 40 cents. I’m crossing a huge bid/offer on the screens (.30/.40) but remember these trades are for educational purposes only. In real life, I would bid at mid, and maybe do a call spread (buy 63/sell 64, etc.)

I want to keep things simple enough, and manageable in 50in50. Precise execution strategy isn’t the point of this write-up, but precise execution strategy is super important in real life.

Conclusion

When you determine that there is an asymmetry around an economic release, you can often come up with and structure positive EV trades. Always understand the following when trading events:

It’s about probabilities. You will be wrong a lot. That’s OK. This FXA trade is probably 50/50, but I think it will pay 2:1.
Don’t do anything unless you are sure you can risk manage the trade. Jump risk and gap risk are scary and they should be scary. Avoid risk of ruin.
Don’t overstay your welcome. If you are correct on an event, the right move is usually to cash out within 24 hours. With this trade, if CPI is weak and AUDUSD goes up, I’ll exit or hedge on the open Friday (14OCT). That will give the trade all day Thursday to get busy and will eliminate weekend risk. This is a ~24-hour time horizon trade.

This post was longer than usual and yet I have left out considerable detail and information in an attempt to keep it manageable. Trading events is risky. It’s fun, scary, profitable, and dangerous. Even if you’re a professional—be careful.

Thank you for reading. I appreciate your time.

If you liked this week’s 50in50, please click the like button. Thanks.

Trade at your own risk. Be smart. Have fun. Call your mom.

DISCLAIMER: Nothing in “50 Trades in 50 Weeks” is investment advice. Do your own research and consult your personal financial advisor. I’m putting out free thoughts for people who want to learn. This is an educational Substack. Trade your own view!