Reading Between the Lines of the Jobless Claims Data

Includes: DIA, QQQ, SPY
by: James Picerno

Today's update on weekly jobless claims is encouraging, but no one should confuse last week's sizable decline in new filings for unemployment benefits with the all-clear signal for the economy. There are still too many risk factors bubbling to convince us that full-bore optimism is now the sentiment of choice.

The Labor Department reported that jobless claims fell to 348,000 for the week through February 9, down from a revised 357,000 the week before. That's a handsome tumble, but as our chart below reminds, the broader picture (as indicated by the black line, which graphs the linear trend) still doesn't look encouraging.

Weekly jobless claims are, of course, a volatile series in the short run. Extreme weather conditions, for instance, can arbitrarily reverse long-run trends in any given week. But if we step back and consider the big picture, it's hard to dismiss the rise in initial jobless claims as anomalous and unrelated to economic conditions of late. Indeed, it's because we know of the pain unfolding in the wider economy that we see the growing jobless claims trend for what it is: a warning sign.

That said, it's important to remember that jobless claims will eventually stop rising. Alas, no one knows in advance when the apex will arrive. Despite what the human mind is inclined to believe, economic trends often start and end for reasons that don't look obvious or logical in real time. After the fact, everything becomes clear, and we all berate ourselves for not seeing what was obvious, albeit in hindsight.

So, taking that into account, how does the current trend compare with history? A subjective, back-of-the-envelope reading of the past offers some perspective on the present rise in jobless claims, which began in January 2006, or so it appears to your editor. Claims fell to 279,000 for the week through January 14, 2006, which marks the previous low for the current cycle. By that measure, we're now in our 25th month of generally rising jobless claims.

The good news is that 25 months looks elderly for upturns in this data series. The last three major increases in jobless claims lasted an average of roughly 20 months. Of course, there's a fair degree of variation in any one cycle. Consider that rising jobless claims lasted a relatively brief 17 months during April 1981-September 1982. Next is the February 1989-March 1991 rise, which lasted 25 months. Brevity returned in the April 2000 - September 2001 increase, which lasted 17 months. Statistically, one could argue that the current rise is due to end. If so, that implies that the economy is due for a rebound in the near future.

Having said that, speculations are always dangerous for the simple reason that forecasting is subject to any number of risks that necessarily haunt looking into the future by analyzing the past. Nonetheless, a bit of historical perspective is useful if only to help frame the question of what comes next. With that in mind, we pose the following: Is the current rise in jobless claims due to end? Or, are we due to suffer an extraordinarily long rise in new filings for jobless benefits?

One reason for answering "yes" to the latter is that when we adjust for magnitude, the current rise in jobless claims still looks like it has legs. Although the trend is two years old, the peak in major upturns in jobless claims has historically come at well above the 400,000 mark. As such, the fact that new filings are currently in the 350,000 range implies that there's more upside left.

Unless, of course, we're in a new era, in which case the old standards no longer apply. That's always a possibility with economic analysis, as history reminds in many, many ways. But until, and if we see some corroborating evidence supporting the new era view, we remain cautious in thinking that salvation is just around the corner.