April 16, 2009
IS THE TREND FINALLY OUR FRIEND?
This morning's news that new claims for jobless benefits fell last week is the best news yet for thinking that the recession has peaked. It's still too soon to break out the champagne, as we'll explain. But for the moment, a collective sigh of relief is in order. Maybe.
As the chart below shows, new filings for jobless benefits tumbled by 53,000—the biggest weekly drop since December. More important is the trend. Since reaching a seasonally adjusted high for this cycle of 674,000 for the week through March 28, new jobless claims have fallen in each of the subsequent two weeks, lowering the total to 610,000 last week. That's still an unmitigated sign of recession, but the recent fall also begs the question: Does the downshift have legs?
This is a critical question because, as we've written, initial jobless claims are a valuable forward-looking indicator for estimating when recessions bottom out. In our March 6 piece, we looked at the historical record and found that initial jobless claims peaked concurrently with, or sometimes ahead of the formal end of recessions since the late-1960s. That's valuable information since identifying the end of the business cycle downturn is much easier after it's obvious to the crowd. The National Bureau of Economic Research, which officially dates the start and end dates of recessions, makes its proclamations long after the fact. Meanwhile, most of the popular metrics for gauging the state of the economic cycle, such as the unemployment rate, are lagging indicators and so they're among the last to reveal when the recession has turned, much less ended.
Initial jobless claims, then, are a better albeit less-than-perfect metric to watch for gauging when the cycle may turn. There are other leading measures to watch as well. Indeed, the stock market's upturn of late has arguably been signaling that the worst of the recession has passed.
But while it's tempting to pronounce the cycle has turned, such thinking is still premature for a number of reasons. That includes the view of some economists that last week's numbers should be ignored because it came during a holiday week following Easter. Meanwhile, the war on deflationary pressures is still raging and key sectors of the economy are still bleeding quite heavily. The latest clues include yesterday's news that consumer prices posted a modest decline in March. Meanwhile, the government advises today that housing starts continue to sink (falling nearly 11% last month vs. February), as did new building permits (down 9% last month), a signal that the outlook for a rebound in construction remains dim.
Let's also recognize that even if the recession has bottomed out, that's a long way from saying that a return to growth is imminent. It's likely that the economy will tread water for several quarters at the least once the economy stops contracting. And while the stock market appears inexpensive, or at least fairly priced, it's still too early to expect that profits are set to rebound any time soon—for reasons we'll be discussing in more detail in the next issue of The Beta Investment Report.
Still, it's not too early to begin elevating risk exposures in those asset classes and their subcategories that are most attractively priced. If we were supremely confident what was coming, we'd be more aggressive in our adjustments to asset allocation. Alas, we're only mortal, and so we continue to act accordingly.
Meantime, we're watching the leading indicators and trying to figure out if the apparent dawn is real or false. Coming to something more than a guess will take a few more weeks, perhaps a few more months. Let's hope it doesn't require several more quarters.
Posted by jp at April 16, 2009 10:53 AM
The drop is for the week through Jan 3, 2009. No doubt there's a seasonal factor there, and one that was unusually large--large enough to overwhelm the normal seasonal adjustment reflected in the graph. As with any one data point, there's always short term noise, and so it's best to focus on the trend.
Posted by: JP at April 21, 2009 8:47 AM
What do you make of that big drop a few months back?
Your chart only has one year marks on the x axis so I can't say exactly where that drop took place. How does that affect the correlation?
Posted by: Tom Verso at April 20, 2009 6:49 PM