Well, if you are looking for accurate numbers to base long(er) term investment decisions on, then yes they are. If you are just trading for the day then it really doesn’t matter. The markets react to these economic data sets as if they were real, so in the short(er) term they matter. But, all of Wall Street knows these numbers are inaccurate at best and more than likely, completely useless. But in a typical the-emperor-has-no-clothes scenario they all feel that they must pretend the data is (somewhat) helpful simply because everyone else does. Now, this is no secret to many PrimeRates readers, but for many people these reported numbers are, for the most part, considered the U.S. Government’s unvarnished best estimates. And they sorta, kinda are, but the way they go about gathering the data is not exactly cutting edge:
First and most important is that the both the monthly Employment and Gross Domestic Product initial releases are based primarily upon surveys not reality. That’s right surveys. In today’s real time world, the US government depends upon sample surveys subject to massive revisions when the actual data arrives months and years later. In other words, both initial jobs and GDP are nothing more than wild ass guesses!
Well, that may be a bit hyperbolic, but the point is well made. The fact is that there are far more accurate (and easier) ways to get some of the economic reports that would require far less work. Taking massive surveys every month depends on a lot of variables, including reaching a sufficient number of people and/or businesses. But even if that was accomplished properly, it is the equivalent of taking a poll. Yes, the data will have a certain closeness to reality minus any margin of error, but inevitably some of the polling will be old news in a fast moving economy. Further, there is actual, hard, up-to-the-minute data available at the government’s fingertips:
And what makes the initial employment numbers an even bigger joke is that there is real time income and employment data available, but totally ignored by the US government. Every time you or I get paid, the amount of withheld income and employment taxes gets sent to the US Treasury and reported the next day online in the Daily Treasury Statement.
That’s right, another part of the government has these numbers and quietly reports them every day. On top of that, the survey numbers that are widely reported use what are called “seasonally adjusted” results. What that means is that they take their survey results and change them to reflect what they see as normal swings in the marketplace. So, if the past indicates that say, July, is a normally slow month in the labor markets they will add in jobs to the final number to “smooth” out the data. Of course, pollsters do the same thing (and the reason many of them were off by, in some cases, quite a few points in the last election). As pollsters have to guess the turnout of various parts of the electorate, they often base it on whatever the last election’s numbers were. This is why many people call polling part art and part science. But it can also lead to misleading polls when guessed wrong, even by a fraction. There are other ways to get real time data as well, but the point is that government numbers are essentially just polls and have limited value. In this day and age there are real numbers available and it is a mystery why it is not relied on more. The article linked is from Trim Tabs and they always show the monthly payroll data (as some other sites do) and invariably that data always seems to be proven more accurate once the “revisions” roll in months later from the surveys.
Why does the government rely on surveys for employment and other economic data? No one seems to really know the answer. Polling (surveys) have their uses but one would think that investors would pay more attention to the actual numbers. For now though, it seems that Wall Street will continue to react to the polling data far more than the actual data. And if you are doing short term trading, that reality cannot be ignored.