Thursday, August 19, 2010

The bounce in weekly claims could be just bad seasonal adjustment

The top chart shows the seasonally adjusted version of first-time claims for unemployment, while the bottom chart shows the actual number of first-time claims. The one that's reported and commented on is the top one, while most people ignore the bottom one.

The story behind today's unexpected rise in claims to 500K is that actual claims didn't fall by as much as the seasonal factors expected. Yes, in the most recent reporting period, claims actually FELL by 22,600, whereas the reported number (seasonally adjusted) showed a rise of 16,000.

I have never believed that week-to-week changes in unemployment claims hold any significant information content, because for one, claims are subject to the vagaries of seasonality and faulty seasonal adjustment factors, and two, the economy rarely changes on a dime, from one week to the next.

Looking at the bottom chart, I can't find any evidence suggesting that the labor market has suddenly deteriorated. Note that the rise in actual claims in early July translated into a big seasonally adjusted decline. In other words, actual claims in July failed to rise by as much as the seasonal factors expected. Well, today's number could be simply the flip side of that number: claims failed to fall by as much as expected. In seasonally adjusted terms, the recent rise in claims could just be "payback" for the July decline. We'll have see what develops over the next few weeks, of course. But I think it's premature to jump to the conclusion that the labor market has suddenly taken a turn for the worse, just as it was premature to think that the big decline in claims in early July was a sign of dramatic improvement in the labor market.

Reading the economic tea leaves is never as simple as watching one series for ups and downs. You have to take into consideration a number of factors, and look for consistent patterns that tie them together. For my money, I think that recent strong growth in commodity prices, coupled with the strong growth in industrial production, strong growth in corporate profits, strong growth in shipping and rail activity, and the ongoing decline in corporate bond yields and spreads, suggests very strongly that the economy on balance is on the mend, albeit relatively slowly. To be sure, construction has yet to really improve, and the labor market is still distressed, but you can never expect everything to move in a straight line and at the same time. For that matter, the labor market is typically the among the last sectors of the economy to participate in a recovery.


狂猪 said...

Does the extension of the emergency unemployment claim have an affect on the initial claims number? Or does the EUC mainly affects continuing claim?

Scott Grannis said...

Emergency unemployment claims are separate from the weekly initial claims and separate from the continuing claims.

Bill said...


Does the weak number from the Phily Fed suggest that growth will slow down if not contract for the second half?

Scott Grannis said...

I haven't commented on this because I consider the Philly Fed data to be too volatile and thus not a reliable indicator of what's really going on in the economy. It has given some seriously false signals in the past. For example:

It was very weak (weaker than today's release) during the summer of 1995. But the economy grew moderately (at about a 2.5% rate) during 1995 with no perceptible slowdown, and then very strongly in subsequent years.

It also fell sharply (more so than the current dip) in Sep. '98, yet the economy remained firmly in the midst of a very strong 4-5% growth rate expansion.

Thus, I think it's a stretch to get exercised about a random drop in this index. Premature to come to any conclusions.

brodero said...

I am glad you noted the use of nonseasonally adjusted jobless claims. I felt in this cycle they
are the data to look at (jobless claims wise) because of some of the
unique structural issues coming on board ( i.e. auto industry,teacher layoffs census workers etc). What we need to see is a break of the
May NSA lows which was 395,000. Today's numbers should have come in below but it didn't which was a disappointment. If next week breaks
that number then we know it was a seasonal adjustment factor issue.

brodero said...

Another use of the nonseasonally
adjusted jobless claims is its 52 week moving average...which in the past has tracked well with the unemployment level and the unemployment rate....

John said...

I have read that for the first time since 1962 the dividend yield on the DJIA is higher than the yield on the 10 yr US Treasury.

IMO the market is pricing another recession into bonds and equities that is not (yet) occuring. The prices of US treasury bonds could be approaching a 'blowoff' stage that reminds me of the technology dotcom blowoff that occured in 1999.
Today's negative reaction to one unemployment number is indicative of kneejerk reactions to headlines quite reminicient of the very late nineties when positive reports of new 'eyeball count' estimates caused near panic buying of securities with little fundamental value.

IMO there is very little value and not a small amount of risk in the longer end of the bond market, and virtually no yield in the shorter end. Equities represent ownership, not just to one year's earnings, but to a stream of earnings and dividends into the indefinite future; earnings and dividends that can increase in a growing global economy.

We do not know when, but sometime soon the emotional momentum of pouring money into the bond market will end and bond investors will be confronted with something they have not experienced in quite a losses. And these losses will likely dwarf the pittance they are recieving in interest income. It is not yet clear which asset classes will benefit from the coming exodus from bonds, but it is a fair guess that companies paying and raising dividends and selling at decade and longer low multiples of earnings and cash flows will be significant beneficiaries. Companies like those in the DJIA.

When (not if) the turn comes it will be fast and furious.

Just my cheap opinion.

brodero said...

If California's initial jobless
claims performed the same way as the rest of the country...the initial jobless claims would be 17,000 less.....California has had
very little improvement in jobless

Benjamin Cole said...


I like your analysis of the bond market. I think there is the potential for a huge, secular bull market in equities and properties, accompanied by long-term growth and very low interest rates.

My worry, as always, is that the Fed actually considers low rates to be a sign of easy money, when Milton Friedman and others have pointed out that low interest rates are actually a sign of tight money--low inflationary expectations.

Interest rates may be low due to abundant capital--high global savings rates. Certainly, investors are willing to buy bonds for nothing.

I think all we need now is an aggressive, expansionist Fed. The doors are wide-open, inflation is very low, in fact we appear to be facing deflation (I think if one could measure all the dealing that goes into making a sale, we in fact are in deflation now, but that is an anecdotal observation).

Maybe it will take another year and I am just impatient. I am hoping for a long, sustained rally equities and property rally, as we withdraw from Iraq and Afghanie, and rein in federal outlays.

It would be nice if Obama and the D-Party shucked their anti-business rhetoric, but the actual regulations and tax increases are marginal--and we are still less regged and taxed than in many previous prosperous decades.

Interesting question: Everybody says they hate regulations. Trucking companies want to take triple-trailers onto federal highways. That's a rig with three trailers behind it.

Okay, less onerous regs anyone? And to be fair, let's say your daughter regularly commutes along a federal highway.

joemckendrick said...

Rosabeth Moss Kanter posted a good explanation for the prevailing blues about the economy: it's a natural tendency to lose faith in the middle stages of a process, as things never seem to be happening "fast" enough.

Posted last year, just as relevant this year:

John said...


CNBC is reporting that beef prices have spiked to their highesr level in two years. This, of course, is in front of the big Labor Day cookout bonanza and it will likely settle down as we move into the fall but IMO it is indicative of how quickly things can change. A drought in some parts of Russia has sent wheat prices leaping and suddenly there is a huge bid for the largest fertilizer company in the world. Now the beef scramble.

Intel is buying McAfee, FOR CASH not stock (shares obviously are too cheap to use as currency). Ditto BHP for Potash. Banks may not be enthusiastic lenders but bond issuance is strong, particularly at the lower grade end as investors seek yield and companies refinance and extend maturities.

Sometimes this takes awhile to build. Monetary easing is known for its long lags and this time is no exception.

One other piece of PURELY anecdotal evidence: I own a VERY, VERY small percentage of a stub end piece of undeveloped commercial property in my old hometown. It has been on the market FOREVER. Guess what? We have an interested (and very qualified) buyer. Money is not happy earning nothing. It is going to seek out returns, and that means real property and other assets that can generate income. Sometimes it happens slower than we would like but if money stays this cheap, the economy will come back strong...we just don't know how long the lag is. I remain convinced we will face inflation, not deflation over the next several years.


Thank you for the heads up re Marc Faber. I am always interested in what he is thinking. I'll try to catch Kudlow's program.

One other thing...Bespoke has thrown cold water on the DJIA yield vs the 10 yr treasury. It was higher in the '08 '09 market bottom than now.....but only slightly. I still say high yield equities are the cheapest asset class out there right now.