Why nailing the jobs report is actually 'impossible'

The market is gearing up for the April jobs report on Friday, which will give an indication of whether the economy is bouncing back in the second quarter, and how many jobs America is creating.

But just how good of an indication will it be?

The Bureau of Labor Statistics' monthly Employment Situation presents numbers compiled from two data sources: the payroll records of some 143,000 businesses and a telephone survey of about 60,000 households. The numbers are then seasonally adjusted. And over the next two months, the BLS rejiggers the numbers "to incorporate additional sample reports and recalculated adjustment factors," which can easily lead to big shifts in the data.

The BLS makes no bones about the precision of its figures. The bureau says the nonfarm payrolls number has a 90 percent confidence interval, meaning there is a 90 percent chance that the actual number of jobs created is within 105,000 jobs of the estimate, in either direction.

If "the reported nonfarm employment rise was 250,000, then… it is likely (at least a 90-percent chance) that nonfarm employment had, in fact, risen that month," the BLS writes in an explanatory note, employing what is in fact a rather striking example.

The employment report is often a big market mover. But there is enough mystery surrounding the widely watched number that the report is basically untradeable, says Vincent Mayeski, an algorithmic trader whose company, TradeXoft, develops trading and market-making algorithms that are used in Treasury markets.


Mayeski previously worked as a quant at Merrill Lynch, where he developed a prediction model for the nonfarm payrolls that takes in some 250 macroeconomic inputs and produces a jobs prediction. He continues to run the model to this day (for April, it spits out a number of 220,000 jobs created, which is very close to the 224,000 Thomson Reuters consensus). However, he elects not to trade on it.

Read MoreWill the jobs report be the spoiler?

"You probably want to have a bias based on a model, but I don't know that you would take a big position, because it is a high-variance, noisy type of a release," he told CNBC. "As much as you can know about the economy and jobs market, there are always seasonal factors that are impossible to predict. So to trade it and make money consistently is very difficult. It's not a get-rich quick strategy, that's for sure."

On the contrary, because he has seen the massive effect that the report can have on the Treasury markets, Mayeski actually refuses to trade ahead of the 8:30 a.m. EDT release. Instead, he will fire up his algorithms at 8:31.

"I don't want to take any sort of a position, because I don't want to introduce that kind of volatility to my p-and-l," he said, referring to his daily statement of profits and losses. "It's the kind of thing that can easily ruin your whole day."

Of course, just because it is noisy doesn't mean the report is useless.

"I'm one of those people who's probably guilty of putting undue emphasis on the jobs number," said Scott Clemons, chief investment strategist at Brown Brothers Harriman. "Any time that everyone focuses so much attention on a single release, it just goes with the territory."

However, Clemons will look to the report to confirm the direction of recent economic releases showing that the economy is bouncing back in the second quarter after a tough Q1. And in the broader context of recent reports, the number can be quite instructive.

"The jobs report has a strong read-through to personal consumption, which tells you how much Americans are going to be spending, and that's 70 percent of the economy right there," he said. "So I think the attention is justified."

Want to be part of the Trading Nation? If you'd like to call into our live Monday show, email your name, number, and question to TradingNation@cnbc.com.