June Employment Report Preview
Friday's employment situation report comes at a critical time for financial markets. The combined efforts of the bearish blogging network, the televised media (What are the chances of a "double dip, Mr. X?), and the technical community (If we close a couple of points below 1040, the next stop is 20% lower), have scared the daylights out of the average investor.
My guess is that Friday's employment report will make matters even worse. Meanwhile, there will be the usual blunders in interpretation.
Let me start with some background on my approach.
Our Own Estimate
Each month we ask the question, "What change in payroll employment would be consistent with other economic data from the same time period (the middle of the prior month)?
This is not a forecast, per se, since we do not posit any causal relationship among these variables. They are all concurrent indicators of economic activity.
- We use the four-week moving average of initial unemployment claims, culminating in the week of the employment survey. This is the best direct indicator of new job losses. This has increased a bit — 464K versus 454K last month.
- We look at the University of Michigan sentiment survey, which we find to be more useful than the Conference Board's sentiment index. Michigan uses a panel, where some families are carried over from month to month. This is a good technique. Sentiment is influenced by employment. When people have lost jobs they are worried, it shows up in sentiment. It is a good concurrent indicator. The Michigan index is now at 76.0, up nicely from 73.6 last month.
- We use the ISM manufacturing index. This value will be published tomorrow. Since I write in the evening, I try to do my employment preview on Wednesday. This allows readers to see my trading posture in advance. When the ISM number is published on Thursday, I do not have all of the components for the model in time for the article. I am assuming a report of 59.0, based upon today's Chicago report. This is the most bullish of our various indicators. If it differs significantly, so will my jobs forecast.
Our long-term record has been very good, especially when compared to the final revised data. This makes sense because our model was derived from the final data. Our approach makes logical sense, because it involves some factors related to jobs lost, and some related to job creation.
Based upon the data, our estimate is a small gain of 6,000 jobs. This should be adjusted by whatever the BLS announces as the census effect, something that I cannot model.
Other Forecasts
It is always interesting to compare the job forecasts from different sources. We follow several because of the widely varying methods they use. A wise interpretation would be to consider all of these disparate sources of information.
All of us are dealing with unusual circumstances because of the census hiring and later layoffs. This is not something that fits nicely into a model. The various forecasts should all be interpreted within that framework.
ADP has proprietary data because of its payroll management business. Looking only at private sector jobs — no government, no census effect, ADP sees a gain of only 13K jobs.
TrimTabs has another valuable approach — tax deposits. Their forecast is a gain of 91,ooo private sector jobs. We then need to adjust for the government shifts. This is particularly noteworthy since TrimTabs has been (correctly) more bearish on job growth in recent months.
Briefing.com cites the consensus as a loss of 100K and their own forecast is a loss of 145K.
WANTED Technologies is no longer doing a public monthly forecast. This is a recognition of the economic realities. My preliminary data (yes, I know I promised a complete accounting) suggests that their innovative method was actually better than the monthly payroll report. They should be rewarded for good work. Instead, they do not find it profitable to publish their findings.
How can this be? The correct way to interpret jobs data is to look at the final result when we know what really happened, and then see who did best. No one actually does this. Instead of viewing the various research approaches as a mutual quest for The Truth, the market deems the BLS preliminary result to be "official."
To summarize briefly, the market wants a prediction of the BLS preliminary estimate — including the 100K confidence interval, the seasonal adjustments, and the benchmark revisions. I am probably the strongest supporter of BLS methodology and integrity, but I still see their approach as only one method out of many.
In fact, I believe that the ADP weakened their approach by attempting to predict the BLS result instead of just reporting the facts. I am disappointed at the loss of the valuable Wanted Technologies approach.
If you really seek The Truth, using the final benchmarked change (as we do) then you may have wider deviations from the initial reports.
Investment Implications
This is an Alice in Wonderland world. From an economic perspective, nothing much has really happened in the past month. The employment situation, as reflected in our indicators, has been about the same for several months. It was not as bullish as some expected last month, and it will not be as bearish as people will report on Friday.
The political spin is a different matter. Many people just read the headline. With the loss of census jobs, the payroll report may have a negative print. How will the market react? Even though this is "old news" you have to expect a bearish reaction. There will also be the regular incorrect and negative twist on the birth/death adjustment.
As usual, the best way to trade the employment report is to be short going in.
A Deeper Understanding
Some typical errors can be expected.
The featured commentators will confuse job creation with net job creation. Why is this important? To pick just one reason, the number of actual jobs lost in the last year is on the order of 30 million. Many more people are touched by unemployment than the net numbers suggest. If you really want to understand employment, you need to start with actual data — not forecasts. Look here.
The failure of the mainstream media and the economic blogosphere to report about employment dynamics is one of the biggest questions I have with the alleged strength of economic blogging.
The big-time bloggers and MSM blogging wannabees are all MIA when it comes to employment dynamics.
This is a key to understanding the negative sentiment. No on is getting this right.
UPDATE, Thursday Morning: The ISM report for June was only 56.2. While this is still consistent with strong economic growth, it is the only really positive element in our forecast. Using this lower number changes our model result to a loss of about 40K jobs, not including the census effect.
Why is the best way to play it as “usual” to be short going in? I just looked at SP numbers over the last few years both on an aggregate and individual year basis. The median 1-day return is higher for every time period when comparing the first Friday of the month to all 1-day returns in the same time period. Additionally, the percent positive is higher under all time periods. The difference in the maximum drop for the day was statistically insignificant for most time periods.
Only the average return for 2010 was lower and this is due solely to last month’s 1-day change. The returns have been negative the last 2 releases.
The time periods I looked at were 2010, 2009, 2008, last 3 years, and last 5 years. I’m assuming that unemployment has been released on the first Friday every month.
Limitations of my quick study:
I just looked at the returns – I did not focus on whether or not the jobs report beat expectations (I didn’t have the data or I would’ve) or any other factors such as what the market did the previous day or if the market was oversold/overbought going into it. Finally, perhaps your statement wasn’t for the entire day just the morning and I do not have the data of minute-by-minute ticks to do that analysis…but would love to see it. Then again, I’m not a day trader.
I’m not trying to be jerk – I really love your blog and respect your knowledge. I’m just trying to understand in case I’m missing something. Thank you for all you do.
If we close a couple of points below 1040, the next stop is 20% lower)
There is a pretty strong support level at 940-950. I think that is the next stop. The next is at 870-880 which in my opinion has monster support. It is also the typical correction magnitude (25-30%) following the rally (+80%) after a market crash (-60%). 870-880 is also an attractive valuation level. For now, I am hedged having bought SDS yesterday when the S&P was still above 1040.
The trend is clearly down, and we are in the period of historical seasonal weakness. History would say to look for a bottom in the Sep/Oct timeframe. IMO, one gets 100% long if and when the S&P hits 870-880 in the Sep/Oct time frame. The average investor seems to me often has no plan and is completely reactionary in the moment.
Andrew — Thanks for sharing your research. I do not have a study to share with you, only my experience over the last few years.
On days when the number is ostensibly good, there is always some kind of selling. CNBC gets an email bombardment from those with objections of one sort or another. The report is so complex and everyone hates the government. It is easy to raise doubts.
That is why there is always a good chance to cover. If the report is bad, the market just sells off.
I have been following this report for over twenty years, in much more detail than most, and I am really trying to share something that a few might find useful.
It works better if you are trading a futures account, since some of the action comes before the stock market opening.
Thanks again for good data, and a great comment.
Jeff
As we would observe the employement activity it’s definitely low, this is still I think the effect of recession.