Last week I got long the US 5-30 year yield spread (Air Says Get Long Steepeners). I thought the seasonality, combined with a bear move in bonds, would cause some steepening in that part of the curve.
Well, I got the bear move in bonds, but not only was there no steepening, the 5-30 spread flattened down to the previous lows of 92 bps!
What the heck is going on? The initial move down to 96 bps could almost be explained by the economic data of the time, and then the climb back up to 100 bps seemed encouraging, but the post FOMC collapse down to 92 bps was peculiar.
If Fed’s FOMC release was hawkish, I could understand the move. And to some extent, that was the first reaction. The US Dollar screamed higher, stocks initially sold off, and bonds got hammered. But if the Fed was so hawkish, the front end of the curve should have borne the brunt. You would logically expect the 2-year to rise the most. Yet the 2-year rose less than the 5’s.
It’s like the market decided the Fed would be tighter, further out in the future. Yet that’s not what the FOMC committee said, nor what Yellen communicated in the press briefing. In fact, there is a decent argument to make that the chances of a nearby tightening increased, but that the total amount of tightening for this cycle decreased.
Let’s have a look at the Fed’s infamous DOT plots which show the FOMC board’s expected interest rate path.
The entire expected interest rate path shifted down! And most importantly, the Fed lowered their terminal fund rate forecast. If anything, this development was dovish. Yet this interpretation is not getting a lot of airplay. Famed ex-Merrill Lynch strategist David Rosenberg is the only one I have seen actively pushing back against the idea that the Fed’s meeting was hawkish.
So why did the 5-year get hammered? If the market thought the Fed was tightening, then they should have wailed on the two’s. And more importantly, if the yield curve is flattening (indicating the Fed is tightening too quickly), why did stocks rip higher later in the day?
And yet if Rosenberg is correct that the Fed was actually much more dovish, shouldn’t that mean a relatively lower 2 and 5-year yield?
The truth of the matter is that the market doesn’t know how to interpret the Fed. Who knows how many of those governors will be there next year. And how much should we really trust the Fed’s forecasts? The reality is that they have proved time and time again that their words mean jack-squat.
In the meantime, the US 5 year note keeps leading the move to the downside. I am obviously wrong (I won’t insult you with “early”). I could handle being wrong if the whole curve was inverting, but it’s like I picked the absolutely weakest part of the curve to own.
I don’t know if this is just post-FOMC noise, or the market’s collective opinion about where Fed policy is headed. If I had to guess, I would say it is not nearly so well thought out. For whatever reason, there is a big 5-year seller, and that part of the curve is offered. Maybe it is all the new issue supply, or maybe some Central Bank out there is stuffed full of US 5’s and trying to pitch their position. For whatever reason, the market is flattening 5-30’s, and my trade has turned into quite the dud.
* * *
Speaking of David Rosenberg, I am pretty sure he was the one who brought the platinum/gold ratio as an indicator for US 10 year yields to my attention, and today’s post jogged my memory enough to dig up the ole’ chart.
From 2009 to 2016, this indicator worked well as a proxy for US 10-year yields. But over the last year, it has failed.
But wait! Didn’t the whole platinum/palladium relationship get messed up with the recent diesel gate and increased popularization of electric vehicles?
Maybe we should change that ratio to palladium/gold. Intrigued, I created a new chart.
Nope, not really what I was looking for. Maybe the relationship is just broken, and whereas in the past the demand for platinum/palladium for catalytic converters (and other industrial uses) versus gold was a signal of economic activity, today it is not applicable.
Yet what if I took the combined average of the two “white” metals versus gold?
Hey, that’s better. Dare I say it, but I think I improved on David’s indicator. Someone let him know…
In the meantime, I now have another model that shows bonds are exactly where they should be. Gee, great…