USDJPY VERSUS GOLD: CORRELATION PART 1

“Hey, the math never lies!”

This is part 1 of some correlation analysis I have done over the Labor Day weekend regarding the relationship between XAUUSD [spot gold] and USDJPY; part 2 will be here tomorrow. First some preliminary thoughts.

Does gold run USDJPY or does USDJPY run gold? On the surface, it would appear obvious that the deeper, more liquid, market of USDJPY would run things; my observations, however, are that it’s exactly reverse. Gold moves much quicker than USDJPY, although both can put in eye-watering moves, and that certainly doesn’t imply one is either slower than the other or faster than the other, cuz at times they both have their “moments”. That said, though, generally speaking it is usually USDJPY that is “safer” in a relative sense than XAUUSD … simply look at how stops are treated in the gold market and I rest my case.

With the Labor Day weekend here in the U.S., I want to start by looking at the long term relationship of USDJPY versus gold over the last approximate 10 years, and then after that, take a look at gold on a short term trading level covering last week.

Since 2008, there have been 3 large moves in gold; 2 bull markets, and 1 horrific bear market. The first bull market started in the week of 10/18/2008, when gold bottomed at 681.40, and lasted to the week of 9/3/2011 when it topped out at 1920.91, for a gain of $1239.51 per Oz. That equals approximately 12,395 PIPS, since 1 PIP in gold is considered $0.10 per Oz. Directly below, the weekly XAUUSD with commentary covering that period.

[Click to enlarge for all charts]

At the same time, covering the same period, USDJPY was collapsing [Yen strengthening] to unprecedented levels in the mid-70’s. USDJPY topped out the week of 10/18/2008 at 102.390, and the low the week of 9/3/2011 was 76.69. That represents a move = -2,570 PIPS. Directly below, the weekly USDJPY with commentary covering that period.

This represents a negative correlation of approximately 21%, which is low, and there are 2 very big reasons for this; 1) the financial world for upwards of 15 years prior to 2008, had been using Yen as the “carry trade” instrument for easy money, and 2) because of this, the Japanese government stepped in and intervened in the market, “jawboned” the market, and did whatever it took from day-to-day to see to it the Yen did not strengthen further. They finally “drew the line in the sand” at the 75 level, and in no uncertain terms let the FX market know that USDJPY was not going to be allowed to go below 75. At the time, some traders didn’t believe them, and learned the very hard lesson of what happens when you fight a central bank. [Umm, you lose.] So, the conclusion to be drawn from this 3 year period, is that it is an aberration in the market due to the BOJ’s aggressive intervention efforts to resist in the strongest way possible to not allow the Yen to strengthen, and not some fundamental shift that says the negative correlation is in danger. It’s simply a historical anomaly.

That brings us to the bear market in gold [bull market in USDJPY] that topped out in the week of 9/3/2011 at 1920.91 and ran all the way to the week of 11/28/2015 at 1046.62, for a loss of $874.49 per Oz., or approximately -8,745 PIPS. Directly below, the weekly gold chart with commentary for that period.

During the same time, USDJPY bottomed the week of 9/3/2011 at 76.69, and topped out the week of 11/28/2015 at 123.669, for a gain of approximately 4,798 PIPS. Directly below, the weekly USDJPY chart with commentary for that period.

Between the 2 markets this correlates to approximately -55%, meaning simply that if gold goes down $1 per Oz. [10 PIPS], USDJPY rallies 5.5 PIPS on average.

This brings us to the start of a new bull market, which is ongoing; this started with gold bottoming in the week of 11/28/2015 at 1046.42 and at the close of trading this last week the high for the week was 1329.04, a gain of $282.62, or approximately 2,862 PIPS. Directly below the weekly chart with commentary for this latest period.

During this period, USDJPY the week of 11/28.2015 topped at 123.669, and the low this last week of 8/27/2017 was at 108.265. That’s a move of – 1540 PIPS. Directly below the weekly chart with commentary for this latest period.

Between the 2 markets this correlates to approximately -54%. So, from the bear market to the new bull market, without the direct intervention efforts of the BOJ, the 2 markets correlate roughly the same at around the 50% level. Ok, this is the longer-term correlation; what about shorter, trading type periods? Do these correlations still hold? SHORT ANSWER: Yes, they do.

I looked at the week just concluded of 8/27/2017. I took M15 intervals from each day of the week [M-F] starting at the European open [7:00 Turnkey server time]to about an hour from the Wall Street close [19:00]. In total, this is 240 M15 time periods for each market; what I measured is the close of each time period for each market, and then took the difference in PIPS from one M15 to the next. I put the data from each market for the exact period, and correlated the data into a “plot graph”. Directly below, the correlation of USDJPY and XAUUSD over short periods.

The red line is the approximate path of linear regression of the data, and the means of each period fall closest to this line. As you can clearly see, the 2 markets negatively correlate along the 2^nd and 4^th quadrants of the Cartesian graph, with a negative slope of approximately 50% [45⁰]. The graph below highlights the general nature of the data according to a very simple linear relationship.

The negative “B” coefficient will make the line move somewhat over many data points, with the linear relationship always going through the “X” & “Y” axis.

There isn’t much data that falls in the 1rst or 3^rd quadrants of the “plot graph”; this would imply a “positive relationship” between USDJPY and gold. The data that does fall into these quadrants is very close to the “X”, “Y” axis at 0; the reason for this is simple. When the markets slow down, it’s not at all unusual to see gold up $0.15 at the close of an M15, and USDJPY up 0.8 PIPS, or something on this order of magnitude. Simple random order flow, and/or the individual dynamics of each market can be in play here over the very short term, and you shouldn’t infer from these instances that the correlation is positive. What you don’t see from the data in the “plot graph”, are data points in the first or third quadrants that are far away from the X,Y=0, because that would most definitely signal positive correlation between the 2 markets.

In general, I’m not crazy about trading correlations in markets; however, having said that, unless the BOJ is pulling strings somewhere, the correlation has a history that is relatively stable around -50%, meaning USDJPY is opposite gold by about half in terms of PIPS, and the correlation holds up short term as well as long term; what’s not to like? Knowing this, and the fact that gold moves a few seconds faster than USDJPY [it’s deeper, with larger volumes that need to be adjusted … that takes a few seconds] is something I’ll explore tomorrow.

Markets are closed today for the Labor Day Holiday, and of course the news of the Norks going thermonuclear has upset overseas markets starting last night [Sunday], at the week’s opening; rather muted I would say given the level of the threat. However, the world has been dealing with this Nork nutjob and his family for almost 60 years now, so the bombastic rhetoric and shows of force are nothing new. Still, you never know what a crazy guy will do, and that has propped gold up and put a ceiling on USDJPY [for the moment].

A very quick selloff in Europe to start the day, but 23 minutes into that and a 40+ PIP drop, and there’s your range for the entire day, as once again it’s “speed of light trading … crickets” thanks for coming. And since then it’s been a mind numbing Holiday atmosphere “Flying Wedge of Death” [FWD], aside from the mini $5 “flash crash” in gold that saw the low of the day there [so far], and sell stops taken [again] to the woodshed for a beating; a move which saw USDJPY do totally zero in response. By the time the first hour is over in Europe … “that’s all folks!” … and there isn’t a damn thing to do but wait for Asia tonight and the start of Tuesday trading. Such are Holidays, so no trading today, and even though there were a couple of algorithm signals, they don’t mean much in context of Holiday trading.

I wanted to get this correlation analysis done before Hurricane Irma hits here in the Caribbean, and with Part 2 tomorrow, more than likely I’ll be posting early, cuz I don’t know how long internet will be up and running either by traditional means or via my phone. At some point it’s all gonna get shut down, and I have no idea when it comes back … a day, 3 days, a week? Down here, nobody knows when anything gets done. To that end, I delayed working on the “Open Office” app, and more than likely the PAMM spreadsheet will be up tomorrow. I think it’s important that the correlation analysis be a part of your trading plan, and while not “officially” part of the algorithm, it’s something to very much keep your eye on as you trade … that’s why I wanted to get it done and posted. Onward & Upward!

Time for the beach … the dog & I are outta here … until tomorrow.

Have a great weekend everybody!

-vegas

OUR TURNKEY FOREX “PAMM/MAM” IS NOW OPEN AND OPERATIONAL; SEE “PAMM/MAM MANAGED MONEY PROGRAM” IN “DOWNLOAD LINKS” SECTION IN RIGHT HAND COLUMN FOR DETAILS [VIEW ONLINE AND/OR DOWNLOAD] AND START YOUR JOURNEY FROM WHERE YOU ARE AT TO “ESCAPE TO SUCCESS”!