Stocks did something they’re not supposed to do this week – namely, they fell.
And while that will invariably be the headline across financial portals this evening and over the weekend (something about “worst week for the S&P since August”), the bigger news is probably bonds.
Whereas people are usually joking when they say “this wasn’t supposed to happen” in the context of stocks falling after an extended rally, some folks would tell you there’s nothing particularly amusing about 10-year yields sitting at just 1.68% three weeks into the new year. That is, lower long-end yields really wasn’t something that was “supposed” to happen, seasonality or no seasonality, rebalancing or no rebalancing.
This was the best week for the long-bond ETF since September.
Why is this bad? Well, “bad” probably isn’t the right term, but it isn’t the best news in the world at a time when a lot of year-ahead outlooks rested implicitly or explicitly on the assumption that global growth was set to inflect for the better amid thawing trade tensions and the lagged effect of 2019’s rate cuts.
In the simplest possible terms, yields were “supposed” to rise (even if gradually), and the curve was “supposed” to keep bear steepening (in a controlled fashion, hopefully) providing validation for Q4’s rampant reflation optimism.
Instead, we’ve seen the opposite of that in 2020. There are myriad reasons why, not the least of which are two left-field events in the assassination of Qassem Soleimani and, now, pandemic worries tied to the Wuhan virus.
Crude’s slump obviously hasn’t helped. “Oil prices certainly aren’t doing yields any favors, with the recent slide in crude weighing on breakevens”, Bloomberg’s Luke Kawa writes, in this week’s edition of The Weekly Fix. “But the majority of the decline in 10-year rates since the start of the year is attributable to tumbling real yields”, he adds. “They even traded below zero on Thursday”.
He also notes that forwards “aren’t sending anywhere near as dire a signal about the economy’s ability to get off the mat” as they were in August when real yields were pushing below zero.
That breakevens haven’t fallen further in the face of crude’s slide could be due in part to a “tail wagging the dog” dynamic in the relationship – i.e., oil isn’t in control. “We have run causality tests between US inflation breakevens and oil (breakevens as a driver of the oil dynamic) and surprisingly, the results reveal that breakevens have been a significant driver of oil recently for 5y and 10y US inflation breakevens”, SocGen’s Sophie Huynh wrote this week.
But irrespective of the cause(s), the bottom line is that we’re now back to levels last seen in early November and, even more disconcerting, the 2s10s has now flattened back inside 20bps.
This could all just be a fleeting affair that will “resolve” itself once the pandemic scare recedes and once we get more information about the state of the global economy. And it’s certainly true that the data we have gotten recently has been, on balance, decent.
But until then, this is going to be a concern for the reflation camp. If you ask Goldman, the virus scare isn’t likely to lead to any kind of sustained plunge in yields. The bank on Friday called this week “similar to price action seen in past such episodes (SARS, H1N1, H7N9), which proved to be short-lived”.
“Past large-scale viral outbreaks suggest a large duration rally on news, followed by sharp bounce back”, the bank goes on to say, adding that “past episodes also suggest the news flow could worsen before it improves even if the current instance proves less severe”. The news flow has, in fact, worsened, with France reporting its first cases on Friday.
If you’re wondering when the trough in economic activity comes, it’s usually one to three months after the outbreak. 10-year yields fell between 20 and 40bps over the course of a month during SARS, H1N1, H7N9 scares, Goldman goes on to say.
The latest edition of BofA’s global fund manager survey showed a net 51% of respondents expect the 3M10Y curve to steepen. That was down from November’s local peak, but still near the highest levels in poll history going back to 2005.