22 Sept 2022: USDJPY 145.90 to 140.36 on Japan's intervention, in pips and dollars

[Fairman]

22 Sept 2022: USDJPY 145.90 to 140.36 on Japan's intervention, in pips and dollars

On Thursday 22 September 2022 the Bank of Japan kept its policy unchanged, and USDJPY pushed up to 145.90, a 24 year high. Later the same day Japan's Ministry of Finance bought yen for the first time since 1998, and the pair fell to about 140.36. The ministry later reported spending 2.8382 trillion yen, roughly $19.7 billion, all of it on that one day.

From 145.90 to 140.36 is 5.54 yen, or 554 pips. One lot is 100,000 units, so each pip is 1,000 yen and the move is 554,000 yen, about $3,947 at 140.36.

Now a long position. Say you bought 0.4 lot at 145.60 with a 25 pip stop at 145.35. At a pip value of about $6.88 per lot, 0.4 lot is $2.75 a pip, so the planned loss is 25 x $2.75 = $69. If the stop is filled at 141.50, a level I picked for the arithmetic, the loss is 410 pips on 0.4 lot, which is 410 x 400 yen = 164,000 yen, about $1,159. That is 16.8 times the planned loss.

There's no warning that can be timed here. Officials had been saying for weeks that they might act, and the market kept trading near 145 anyway.

Size the position for the gap you could suffer, not for the stop you set.

[Fairman]

The day didn't end at the low. Reuters had the dollar at 142.31 against the yen at 18:21 GMT, after trading near 140.36. That's 195 pips back off the low, or about 35% of the 554 pip fall. So a short taken at the bottom of the move after seeing it happen was a worse trade than it looked. Someone who sold at 140.50 hoping for another leg would have been under water by 181 pips at 142.31, which on 0.4 lot at roughly $2.82 a pip is about $510.

[Fairman]

Here is a way to size for the gap, with numbers. Suppose you accept a loss of 1% of a $10,000 account, $100, if the stop gets skipped by 400 pips. A pip on one lot of USDJPY at 141.50 is 1,000 / 141.5 = $7.07, so 400 pips on one lot is $2,827. For $100 you can hold only 0.035 lots. Rounded down, that's 0.03 lots. If instead you size from the 25 pip stop, 0.4 lots comes out. The difference is more than tenfold, and it shows how much of normal sizing assumes the stop is going to be honoured.

[Fairman]

When did people learn the intervention had taken place? The market inferred it from the move. The ministry confirmed it that evening, the monthly total came on 30 September, and the daily breakdown, which showed it was all on the 22nd, arrived in early November. In the minutes after the drop a trader had only the chart, which means any decision in that window was a decision under uncertainty about whether it would repeat. That uncertainty is the reason to stay small or flat, not a reason to guess the next official statement.

[Fairman]

A practical checklist for yen pairs when officials are talking about action. First, check the distance to the last major high, since a spike from there is where moves tend to start. Second, cut position size to the gap loss you can tolerate, as in the 0.03 lot example. Third, avoid leaving positions open through the Tokyo close if you can't watch them. Fourth, keep spreads in mind, since a spread that widens from 1 pip to 5 pips on 0.4 lot costs $11 extra just to exit. None of that predicts anything. It limits what the unpredictable event can do.