Tim Duy's Fed Watch: Not Rate Cut News

Tim makes it clear that he is not a fan of the ADP report as he interprets the latest employment data:

Not Rate Cut News, by Tim Duy: If you traded on Wednesday’s ADP report, this is not exactly Good Friday for you. Please, please remember this the next time you feel the urge to pay to much attention to the ADP report. Yes, over time, I believe the ADP number will prove to be not significantly different from the BLS private NFP numbers (which begs the question of why we need two identical measures of the same thing in the first place). Indeed, I find the analysis of ADP’s payroll data to be academically intriguing. And the ADP report may even edge the market consensus in the right direction on occasion.

Yet still, even after the significant methodological improvements made after the December fiasco, when ADP is wrong, it is very wrong from a trading perspective. While the difference between the ADP read and the BLS read on private NFP is “only” 51k, the smallest drop in the bucket of the US labor pool, it can mean the world to the bond markets. And suppose that the BLS number is revised down to the ADP number. To be sure, this will be seen as a victory for the ADP report, but it doesn’t change the fact that you were on the wrong side of the trade today. Simply put, use the ADP number at your own risk, and expect to be blindsided on the first Friday of every month.

The employment report stands against the softer tone of much of the data. NFP rose by 180k, while the prior month was revised up to 113k. For the quarter, the average monthly gain was 152k. I know, if viewed in the light of the Clinton years, this is a disappointment. But the Fed thinks labor force growth is slowing, and consequently is looking for something closer to 100k to hold unemployment in place. Recent figures remain well above that mark; unemployment edged down to 4.4% in March.

In the details, the 56k gain in construction was surprising, but this is being written off as a rebound from February’s weather related decline. Manufacturing employment dropped; no surprise there. On a weak note, business and professional services lost 7k, largely on the decline in the employment services component (but not the temporary help component, which only slid by 800 jobs). Something to keep an eye on. Note also the consistent gains in computer systems designs, up 7.1k for the month and 62.9 for the year – interesting considering there is so much commentary about information technology jobs being sent overseas. And retail hiring is on the upswing – someone must be shopping, despite continue calls for a consumer collapse.

To be sure, we will soon be reminded by the bears that the employment report is the ultimate lagging indicator. I don’t dispute this point, but just look at the last five reads on GDP growth:

2005:Q4	2006:Q1	2006:Q1	2006:Q1	2006:Q1
1.8	5.6	2.6	2.0	2.5

With the exception of the 1Q06 GDP pop, you can’t exactly say that the slowdown is a new event, especially considering that no one believes that 1Q07 was anything special. If we are significantly below potential, I would have expected labor markets to be considerably softer by this point in the cycle. Not only are they not soft, but they are strong enough to generate wages gains – note that the 6 cent gain translates to an annual rate of 4.2%, which will hopefully be ahead of inflation. "Hopefully" meaning it might imply that we are shifting income from corporate profits to labor. Of course, the wage gains could be of the more inflationary variety…let’s not go there just yet.

Still, from a policy perspective, that gets us back to the productivity question (also raised by Jim Hamilton this morning). Note that aggregate hours worked rose at a 1.5% annualized rate for the first quarter – if GDP posts at 2%, you really can’t expect much good news on the productivity front. It was already clear that monetary policymakers believed that labor force growth was slowing. Multiple quarters of soft GDP growth – assuming 1Q07 is weak, 5 of the last 6 quarters will post at 2.6% or lower – combined with continued job growth and a declining unemployment rate, is going to put going to raise some serious questions on Constitution Avenue. Those looking for a rate cut on the back of weak growth should be looking closely at the possibility that relative to potential, growth is not that slow. The Fed will be. In this environment, some good news on inflation would be very welcome.

The risk of a recession is still out there, despite the employment numbers. All the things the bears say will happen may still happen. But the recession keeps being put off to the next quarter. Eventually it will happen. But timing is everything.

Comments

says...

I think you're too hard on the ADP number:
Here's what I saw:
This month:
ADP nailed manufacturing (-16 vs. -15) and services (+128 vs +137)
The source of the miss was clearly the as you note "surprising" construction number due to unusually cold weather in Feb and unusually warm in Mar was according to one Street analyst, the third biggest monthly change in the "not at work due to weather" question since 1980.

ADP also seemed to point in the right direction in contrast to the Services ISM.

For Feb, ADP revised its private number up to 65 and the NFP private number was revised up to 69.

In a game where the standard error of the consensus is in the vicinity of 90K, A miss of 1 and and 51 in the first two months of the new and improved ADP sure looks good from here.

[knzn]

I’m not sure what to make of this employment report. Nearly 100% of the net gain in employment can be accounted for by five categories: construction, education (including local government), health care, general merchandise stores, and restaurants & bars.

[Tim Duy]

Regarding the ADP numbers, I do believe they will over time yield a result very similar to the official numbers, if not virtually identical. And they may often pull the forecast in the right direction. But the benefits of trading on ADP need to be taken in context of the risks. For example, some of the underlying detail may be dead on, but if you are basing your position on the headline, then does the detail matter?

knzn has a good point about the unusual clumping of jobs. Diffusion of gains is down from last year. Health care I expect; the others....? Of course, any given month contains its own oddities. And the 180k headline needs to be taken in context of last month's low reading. But the average NFP growth over the past three months is solid. And on the household side, the number of unemployed fell by 141k, while employed gained 335k. Good numbers as well.

[spencer]

The big gain in retail is really strange. My measure of retail productivity is slowing significantly and the 0.6% jump in retail hours works implies that March retail productivity -- real retail sales / retail hours worked -- will also be weak.

Given that the employment data does not really change the consensus expectation for 1st Q real GDP growth what it does is sharply reduce the implied productivity growth numbers.

On a smoothed basis the data does not change the story that employment and wage growth is peaking and given the recent volatility in the data the use of smoothed data clearly seems prudent.

[calmo]

Is the clumpiness to be explained by the (somewhat crude, possibly suspect, and revisable in any case) use of the birth/death model?
Now about those exiting Mexicans...leaving room for 56k construction domestics to get in there and pound nails at only twice the price.
No, that wasn't it: it's about the extra retail help needed to service all those Mexicans looking for souvenirs...in advance of the summer avalanche of vacationers who need to forget how crummy their financial picture has suddenly become.
Ok, has the BLS been adequately funded lately?