March Employment Report Preview
Fasten your seat belts!
We are about to see something really new — the release of the monthly jobs report on a day when the market is not trading.
Some think this is wrong.
Jim Cramer, with an assist from Larry Kudlow, tried to turn this into a religious crusade. Somehow the decision to release data according to the regular schedule was portrayed as an affront to those observing Good Friday. Here is the link and the video:
Cramer and Kudlow are completely wrong! The US government is not on vacation tomorrow. In this country there is no establishment of a religion, and therefore normal government functions are not geared to religious holidays.
Cramer loves these rants where he gets to lecture to the government officials who really make decisions. Here is the reality:
- The employment report delivers information to the entire nation, not just the financial markets. Message to Jim: It is not all about you — or the markets. Most people are interested in what is happening in jobs and the economy, not what markets do.
- Why should the government adjust to the NYSE schedule? Who elected them as the arbiter of what is a holiday? This is a can of worms for the government.
- If you had a complaint to make, why did you wait until this week? The announcement schedule has been known for months. Why criticize the government, the SEC, and the NYSE now. Look in the mirror.
Government has never paid much attention to financial markets in making announcements, and this is no exception. It is business as usual.
I'll suggest more about how to trade this in the conclusion. For now, we have extra implied volatility. Making a small adjustment from last month….
Options traders distinguish between actual volatility, recorded and measured from data, and implied volatility. The latter term comes from solving the options models using price and the various known inputs. Typically volatility is the only input term that must be estimated. High implied volatility means that traders of options are incorporating uncertainty in their estimates. With this in mind…
Get ready for maximum potential volatility in a single weekend!
The Data
We all want to know whether the economy is improving and, if so, by how much. Employment is the key metric since it is fundamental for consumption, corporate profits, tax revenues, deficit reduction, and financial markets.
We would like to know the net addition of jobs in the month of March.
To provide an estimate of monthly job changes the BLS has a complex methodology that includes the following steps:
- An initial report of a survey of establishments. Even if the survey sample was perfect (and we all know that it is not) and the response rate was 100% (which it is not) the sampling error alone for a 90% confidence interval is +/- 100K jobs.
- The report is revised to reflect additional responses over the next two months.
- There is an adjustment to account for job creation — much maligned and misunderstood by nearly everyone.
- The final data are benchmarked against the state employment data every year. This usually shows that the overall process was very good, but it led to major downward adjustments at the time of the recession. More recently, the BLS estimates have been too low. (See my prior preview for a more detailed account of this, along with supporting data).
I think the BLS is honest and does a good job, which seems to put me in a small minority of observers. Despite this support, I question the general concept. The BLS tries to estimate total employment in one month, total employment in another, and subtract the two to determine the difference. When you are talking total payroll employment of over 130 million jobs, even small errors are in the range of 100K jobs or more. Meanwhile, smaller discrepancies from expectations are unwisely viewed as significant.
Competing Estimates
The BLS report is really an initial estimate, not the ultimate answer. What we are all looking for is information about job growth. There are several competing sources using different methods and with different answers.
- ADP has actual, real-time data from firms that use their services. The firms are not completely representative of the entire universe, but it is a different and interesting source. ADP reports gains of 209K private jobs. Steven Hansen at Global Economic Intersection endorses the ADP method over the BLS result. He has a strong analysis covering many nuances in the data. For those who really want to understand the jobs story, it is well worth reading.
- TrimTabs looks at income tax withholding data. The idea is that this is the best current method for determining real job growth. TrimTabs forecasts gains of about 187,000. They have been pretty bearish on job growth. (Their real-time estimate was for a gain of only 45,000, but they came forward and admitted a "calendar quirk" that threw off their forecast. There have been a number of changes in the tax withholding rules, so let's cut them a little slack and give credit for being honest about what happened and their methodology. They provide a useful additional method).
- Economic correlations. Most Wall Street economists use a method that employs data from various inputs, sometimes including ADP (which I think is cheating — you should make an independent estimate). I use the four-week moving average of initial claims, the ISM manufacturing index, and the University of Michigan sentiment index. I do this to embrace both job creation (running at over 2.5 million jobs per month) and job destruction (running at about 2.3 million jobs per month). In mid-year the sentiment index started reflecting gas prices and the debt ceiling debate rather than broader concerns. When you know there is a problem with an input variable, you need to review the model. For the moment, the Jeff model is on the sidelines.
- Briefing.com cites the consensus estimate as 215K and their own forecast is for 250K.
A Special Message for Maria
In an interview today with David Malpass, Maria Bartiromo acted like monthly net job creation of 300,000 jobs was an impossible dream. I wish that I had five minutes to talk with her about employment dynamics. If we could increase job creation by 10% (without losing jobs) we would add an additional net job gain of 200-250K jobs per month. That would be a net job gain of 400K or so. No on in the mainstream media seem to understand this.
Trading Implications
What does this mean for our trading and investing?
My basic conclusion is that it is great for our enhanced yield program. A lot of implied volatility will come out of the market by Monday.
Let us suppose that (like Mr. Beeks) you knew the jobs report in advance. How would you trade it?
Investors would like to see economic strength — a strong report.
Traders have this irrational fixation on more QE from the Fed. The current mantra, made especially obvious by this week's action, is that any perception of more QE is good.
From this perspective, a good report will be bad and a bad report good!
This is why I am not "trading" the report — just collecting options premiums in a sideways market.
Ben is warming up the presses as we speak…
A very good post regarding the BLS and job stats. Thanks.
Well written Jeff! as always a helpful analysis of market moving events. Have a great weekend!
Jeff, I disagree with your argument about the release date. Yes the employment data is for everyone, not just financial markets. But financial markets are the ones that are most time sensitive.
Suppose a retail investor was worried about the ECRI recession call. A release like today with weakness in employment, hours, temp help, retail employment etc… might be interpreted as increasing the probability that ECRI is going to be right.
But, most retail investors do not trade in the futures market and definitely not in the bond market, so they are stuck until monday. The big institutions get to move right away.
Canada got their labour force data out on thursday and I don’t understand why the BLS couldn’t do the same.
******************************
On an unrelated point about CES methodology. Jeff, your description:
“The BLS tries to estimate total employment in one month, total employment in another, and subtract the two to determine the difference.”
is a bit misleading about what the CES actually does. As you know (since you have written about benchmark revisions), the monthly level the BLS estimates is just a ratio which is then applied to the benchmark.
Jeff, maybe you have some ideas, but I think that if the BLS were to design a survey to estimate the change in employment every month, it would still look very much like the CES (maybe with less sample rotation)…
cd7 —
Good one:)
It does seem that he was on the right track with his speech — employment brighter than the rest of the economic data.
Jeff
Angel — I’ll split this into two responses so that we can all discuss separately if there is more to say.
On the timing of the release — I agree that it is important information. Maybe the NYSE should adjust its holiday schedule to conform to the releases, which are known in advance. You also did not respond to my point that any such appeal could have been made long ago.
As to individual investors, on a normal NFP day they get the number and then stocks open an hour later. There is often a gap, so the investor is confronted with dramatically different prices. I have a specific question for you:
Assuming that someone (unwisely) believes that he or she should sell stocks based upon this data point, what difference does it make whether they do it an hour later or a couple of days later?
I’ll elaborate on the “unwisely” and the ECRI in my ongoing recession discussion, but not here.
Jeff
Angel — on the BLS methodology —
I don’t really understand your point about ratios. They do a survey of establishments and then a number of adjustments. It is not really a ratio. They make revisions as more businesses reply, as they change seasonal adjustments, and as they get final data from state unemployment offices.
As to measuring the change in employment each month — they could ask that question, but it would still be a survey. They would still have to deal with firms that went out of business.
I have written very extensively on all of these topics.
This is why I look looking at the various alternative measures along with the BLS data — something that I have advocated for a long time.
Sorry I can’t go better with the answer.
Jeff
Jeff, with the release date, I guess we will just have to agree to disagree, but if Statcan can get the data out on thursday, how come the BLS can’t?
With the high volume ETFs, an individual investor can trade after hours, so they don’t have to wait until 9:30.
As to the wisdom of doing this, maybe it is a mistake to attempt to vary the asset allocation depending on where you think you are in the cycle, but many investors do it.
Jeff, I agree with looking at many different series rather than just the CES. My particular favourite is the CPS wage and salaried for non-farm. Even thought it has a large sampling error, it does not have the complexities that say Trimtabs has with changes in witholding, or ADP estimating from the changes in employment of a self selected sample of clients, or model builders like Jeff Miller using data like the Michigan Survey, where there are exogenous shocks that change the impact of the data. All of these measures has some value, the trick is to weight them appropriately.
If one had multiple measures of the same target variable, say, change in employment, with no “non-sampling” error, the statistically optimal approach would be some sort of weighted average of all the measures, with the weights inversely proportional to sampling error of the measure.
However, non-sampling error is a major issue for trimtabs, adp, modellers, and, i would argue the CES. I think the various measures could still be combined in a weighted average, but one would want to weight inversely proportional to non-sampling error as well as sampling error.
In my view the CPS has the least “moving parts” of all the estimates so it has the least likelihood of non-sampling error. Thus i weight it highly, even though it has the largest sampling error of any of the measures.
****************************
On the CES, for employment it actually is a ratio as is explained in this BLS technical note.
http://www.bls.gov/web/empsit/cestntab.htm#tab2a
the monthly employment is estimated via ratio from the previous month, and ultimately by a chain of ratios back to the benchmark.
Jeff, I’m looking forward to your future postings on the ECRI and a possible recession. I think some of the recent data, including this release, has significantly increased the possibility of ECRI’s recession call being correct, but i’m not sold yet.