New claims for unemployment benefits in the US increased last week, but the trend remains encouraging. The consensus forecast told us that claims would post a modest dip to a seasonally adjusted 270,000 in today’s release for the week through May 23; instead, filings jumped to 282,000, the US Labor Department reports. But the four-week average, which is close to a 15-year low, barely budged and the year-over-year change continues to fall at a healthy rate. In short, there’s not much news in today’s update, which means that this leading indicator continues to point to ongoing growth for the labor market.
On a side note, it’s reasonable to wonder if the decline in jobless claims are at or near a cyclical trough. Even under the best of macro circumstances, there’ll always be a certain number of layoffs as workers transition to new jobs, companies rise and fall, etc. Note that jobless claims have rarely dipped below the 300,000 mark in seasonally adjusted terms for very long over the past 40 years. As such, the sight of claims bouncing around at current levels for an extended period wouldn’t be surprising or bearish. In other words, the big move down for claims is behind us—we’re not going to 200,000.
That said, the data still offers an upbeat outlook for the labor market. Claims dropped a bit more than 8% last week vs. the year-earlier level. That’s a softer decline than we’ve seen recently, but it’s still a healthy fade. The bottom line: if you’re trying to rationalize a bearish outlook on the US economy, you won’t find much support in jobless claims figures. Sure, other metrics paint a darker view. But for now, the numbers for payrolls and claims remain positive.
The worst you can say about claims is that the recent slide anticipates the end of the cycle. History advises that recessions begin at some point after jobless claims make a cyclical low. Perhaps that cyclical low was the dip to 262,000 for the week through April 25. But bears shouldn’t get too excited just yet. The gap between the cyclical trough and the start of a new downturn can last years. In the previous trough ahead of the Great Recession, jobless claims touched a cyclical low of 282,000 in late January 2006—three years ahead of the next recession. It’s anyone’s guess when the next downturn will start, or if we’ve even seen the lows for the current cycle.
Meantime, the outlook from a claims perspective remains encouraging. “The firing side of the equation has pretty much run its course in terms of recovering to normal,” observes Jacob Oubina, an economist at RBC Capital Markets. “The unemployment rate should continue to decline and wage pressures on the back of that should firm up.”
Not a huge point, but the last recession started Dec 2007, so it would have been a touch less than two years after the cyclical low you mentioned in Jan 2006.
Ali,
Yes and no. This is splitting hairs, but since you brought it up… There are several variations on business cycle dating (see the “Notes” section here, at the St Louis Fed site’s cycle date history page: http://research.stlouisfed.org/fred2/series/USREC). Many analysts agree with you and date the start of the Great Recession, for instance, as Dec 2007 (if we’re using monthly data). But Dec 2007 is also the peak month for the expansion that ended, as per the NBER’s labeling (http://nber.org/cycles/dec2008.html) and data (http://nber.org/cycles/cyclesmain.html). The issue, of course, is deciding when exactly an expansion ends and the recession begins. In theory, there’s a precise day of the turning point and it obviously occurs within a given month (leaving aside the possibility of an end-of-month peak). Let’s say for the sake of argument that the expansion ended on Dec 15. As such, the recession began on Dec 16. One can then argue that Dec 2007 was a month that included the peak and the trough. In fact, no one really knows the day of such things; monthly estimates are hard enough. For modeling purposes, my system of rules is to date the recession as the month after the NBER’s peak month, which in this case is Dec 2007. Why use that rule? It’s a clean break that IDs the month AFTER the NBER-labeled peak month. By that standard, the first full month of the trough period was Jan 2008. To be fair, the convention in the press is to cite Dec 2007 as the start of the recession. That’s fine. As long as you’re consistent in whatever rule set you choose (and understand how that can conflict with alternative rules for dating), all is fine.
–JP