Cross border fund flows / fx correlations

March 4th, 2021, 05:04 PM

Foreign central banks have always been major participants in the US Treasury markets, and some of these investors base their investment decisions in large part on spot and forward exchange rates. China, for example, regularly buys dollars (and therefore Treasuries) to help sterilize their current account, while Japan has also historically shown much stronger appetite for Treasuries when the dollar cheapens vs the Yen. In fact, in many such instances, it's more about the currency than the outright level of interest rates that drives these investment decisions. And then of course there are the sovereign wealth funds, the macro funds that play the carry trade, and so on...

The point of all this is that dollar exchange rates can and do play a significant role in overseas demand for Treasuries. Just look at what rates have done since the January 4 highs in both EUR and JPY. At some point, US Dollar short covering will have run its course and the outlook for Treasuries might start to look more constructive again, but we have a big refunding coming up and the few remaining Treasury longs out there are probably feeling pretty nervous here.

Anyhow I'd really love to hear other traders' thoughts on this dynamic, as I'm finding myself once again looking to incorporate exchange rates into my own Treasury trading in a much more methodical fashion than I have in the past. Thinking about whether it makes sense to try to build a model or alert system, and if so, what time frames would make the most sense. Gut says daily / weekly / monthly but it sure would be nice if these inputs were meaningful on intraday timeframes as well.

Again, any thoughts, ideas, experiences, observations, or comments are welcome. Hoping that an open discussion of this topic might stir up some ideas

March 8th, 2021, 07:41 AM

OK I guess I'm on my own here

For what it's worth I tinkered around with some data over the weekend and found that on days that the Dollar was up by at least 0.15% vs. the Yen (i.e. a decline of 0.15% in quoted price of the J6 futures contract, which was roughly 35% of all observations), TY was down about 75% of the time, and when the dollar sold off vs. the Yen by at least 0.15% (also about 35% of observations), TY was up about 74% of the time. By increasing the percent change threshold on the currency the total number of observations declined but correlation was even higher. My dataset included a total of just under 1400 days.

Now these general correlations are already pretty well known so this is certainly nothing new. I'm just trying to come up with ways to better incorporate these concepts into my trading. Is there a lead / lag that can be exploited? Is there anything predictive about the days when the correlation breaks down? Maybe there's potential for a futures spread mean reversion strategy between these two contracts. Lots of different directions one can go with this. Could keep me busy for a little while...

Cross border fund flows / fx correlations

Discussion in Treasury Notes and Bonds

Cross border fund flows / fx correlations