Just Roll With It?
by Brian Pretti CFA, Contrary Investor. August 31, 2007
There has been much written over the past month about credit market turmoil in the US, that has now clearly spread well beyond US borders in terms of fallout. As you might know, I've written about the potential for significant credit market problems for years now, to the point where I know I've sounded like the guy continuously crying wolf. So now the wolf has at long last arrived on scene. Whaddya know? Whether it’s nominal and relative levels of credit market leverage set against historical experience, the absolute mushroom cloud explosion in OTC derivatives, etc., I've tried to cover the waterfront of factual information regarding credit market excess. So rather than joining what has now become a consensus chorus of concern about the credit markets from the very same folks who've had their collective heads buried in the sand over prior years, let's instead look ahead a bit to what may be important consequences to come. Let's face it, we all know what is because it has been as plain as day in the data for years. What we don't know is future fallout.
There's no question that the financial markets over the past month have been dealing with and attempting to discount credit market issues de jour. But the fallout consequences on the real economy may be very meaningful looking ahead. Certainly further deterioration in an already down and out housing market is one huge question mark, especially given the fact that the greatest number of ARM resets we're probably ever going to see in our lifetimes will occur between September of this year through the summer of next. But the next big issue the financial markets are going to have to deal with is the potential for recession. That has not yet been discounted in financial asset prices. Rather than speculate and guess, I thought it timely to home in on just one key economic aspect leading toward the potential for recession that will be more than well watched as we move forward – payroll employment. As you've probably heard in consensus commentary lately, many a pundit has stated, "there isn't going to be a recession as long as unemployment is low and payroll growth continues." In deference to these consensus thinkers, let's have a look at what have been important leading indicators for employment trends and the "messages" they are currently sending. For without question, if you don't believe trouble in mortgage credit markets will impact consumer spending, maybe trouble in payroll employment will. Let's get right to it.
The first stop on this journey is the very meaningful correlation between temp employment numbers and headline payroll trends. THE major importance of this relationship is that temp employment trends lead the headline payroll employment numbers and rate of change trends. History is more than clear on the subject. Have a look.
Over the past decade and one half, temp employment has both peaked and troughed ahead of the headline payroll trend. As such, it has been a very important indicator to follow. By the way, as you can see in the chart above, over the last few months, the year over year change in temp employment has gone negative. Not a good thing in terms of foreshadowing forward headline payroll employment trends. As per historical experience, negative rate of change trends in temp employment occurred just prior to the early 1990's and 2001 recession, but not during the mid-cycle economic slowdown of the mid-1990's. Are current temp employment numbers and trends telling us the next recession is simply not far off? We'll just have to see what happens ahead.
Another very important leading indicator for macro payroll employment trends is retail sales. But unlike its temp employment counterpart, which tends to lead changes in payroll trends both on the upside and the downside, retail sales lead payrolls directionally on the downside. In terms of leading or lagging tendencies on the upside, it's a mixed bag as you'll see in the short history below. But for our current purposes in the present cycle, we're interested in leading tendencies on the downside, so watching the year over year annual average change in retail trends is quite important immediately ahead. If the rate of change in retail bottoms prior to a bottom in annual payroll growth, we can sit up and acknowledge this fact as a potential precursor to a change in labor market conditions for the better. But that's not what's happening, quite the opposite.
Another lead/lag indicator for headline payroll trends just happens to be the trend in the household survey of payroll employment. For those unaware, each month the BLS (Bureau of Labor Stats) really conducts two payroll employment studies. One is the headline (as reported in the financial press) report whereby the BLS seasonally adjusts data and estimates new business formations and hiring. Since the headline payroll report is really based quantitatively on payroll tax records, and new businesses file these with a lag, the BLS employs the infamous birth/death model to come up with the headline number. The second survey, if you will, is the household survey whereby the BLS literally polls folks on the phone. As such, this survey picks up new business formation and hiring trends before the institutional headline report. And especially meaningful in the current cycle is that the household survey can pick up changes in immigrant construction jobs. Again the leading tendencies here are very short term in nature. For now, the year over year rate of change in the household survey is below that of the headline, implying more headline payroll experience downside to come. The important issue moving forward is to watch for a turn or divergence in the household survey relative to the headline numbers and trends. For now, this says more southern journey to come.
What is very important to notice in the above chart is that the rate of change in the household survey has turned up prior to the headline survey trend at each meaningful bottom in total payroll employment. Current message? No bottom in sight for now.
As a brief tangent, through July, headline payroll growth in the US has totaled 955 thousand new jobs. Of this, 773 thousand of these newly created jobs are as a result of birth/death model estimates – 81% of the total. I'm not going to suggest the B/D model is some type of conspiracy by any means, but rather point out what the importance of seasonality in these estimates as we look ahead over the remainder of 2007. Below I've created a seasonality chart for the birth/death model using data from 2000 to present. Of course what is important is obvious. The B/D model is consistently strong during the February through June period each year. After that, its influence trails off rather markedly, with the exception of August. On an average seasonal basis, the B/D estimation tailwind is about to subside in a big way over the second half of this year. In conjunction with a slow economy of the moment, and the fact that so many indicators discussed above continue to point south for forward payroll growth, is the headline payroll report headed for tough sledding in the second half of the year? We'll see.
Last point to watch is corporate earnings growth. In the following graph we're using the pre-tax corporate profit numbers that come to us courtesy of each GDP report from the government. As is clear, important peaks in the year over year rate of change in pretax corporate profits have indeed foreshadowed important rate of change peaks in payroll employment. Like the headline payroll trend relationship with temp employment, the change in corporate profits leads payroll employment trends. Hence, important to monitor.
Although not marked, it’s also true that year over year bottoms in corporate profit growth have lead rate of change bottoms in payroll experience.
There you have it. I believe that the real story of forward US payroll employment growth is being foreshadowed and foretold in the leading indicators for headline payrolls that are temp employment trends, the rate of change in corporate profits, the annual growth in retail sales, and the current downside corroboration in the household employment survey. Of course not once in this discussion did I mention the very large potential for both forward construction and financial services job losses that seem quite clear on the current horizon. For now, many expecting the US to bypass an official recession point to payroll strength as a key rationale for this viewpoint supportive of consumer spending. But the facts show us that tried and true leading indicators of payroll trends are pointing in exactly the opposite direction of strength. For now, the year over year rate of change in payroll growth is below the lows seen in the mid-cycle economic slowdowns of the mid 1980's and 1990's. Historical precedent is telling us to be very suspect of those expecting US payroll growth to remain strong.
I'll leave you with one last question. If indeed US payroll employment growth slows meaningfully ahead, as seems to be the case from the factual data review above, can we really believe the consumer spending and credit acceleration dependent US economy won't blink? Stop listening to the talking heads and other assorted Street fortunetellers guessing about the direction of the US economy ahead and start looking at the factual data right in front of our faces. The real economic data may not be as entertaining, colorful or as hilarious as the circus clowns on CNBC, but I guarantee you the admission ticket for the investment entertainment show can be oh so costly. Remember, once you've purchased a ticket to the television Street pundit/circus barker show of life, there are no refunds.
© 2007 Brian Pretti