GDP, BLS Numbers Are Worthless

The Bureau of Labor Statistics reported 204,000 new November jobs and everybody said wow, the US economy must be doing better then we thought. Similarly, the Bureau of Economic Analysis reported that the US Gross Domestic Product grew by 2.8 percent in the September quarter. If you only used government data, the impression is that maybe the economy is indeed getting better and we can grow our way out of the current no growth swamp.

What nonsense. No one, at least no one I have noticed or than me, has said that maybe the numbers suck. Look, what’s really going on is that the US economy is barely growing and what growth there is at a slower pace than last year. Rather than improving, flat lining is the best description of the US economy.

How do I know this? What I track is daily withhold income and employment taxes reported by the US Treasury and then adjust it for changes in tax rates, salary levels and tax bracket creep. Unfortunately for you guys, no one else I know tracks this and since TrimTabs sells this data as part of our research products, we cannot give it away. But I can tell you that 2012’s modest 3 to 4 percent nominal growth in wages and salaries before inflation has dropped to an average of 2 to 3 percent this year. Why the decline? Simple; higher employment and income taxes have reduced the deficit a bit and also income growth.

Ok, why the initial job and GDP reports suck is that they both ignore all the real time data available and instead are based upon surveys adjusted by historic data; a methodology first developed in the 1960’s. So, let me be clear, about two years after the initial GDP and jobs reports the final numbers will be accurate. But that will only happen after all income tax returns and actual sales numbers are collected. What is most important is that historically the final numbers almost always have no relationship with the initial reports.

Now I know many of your eyes glaze over when you see another video on boring GDP and jobs numbers. But the truth is one of the major problems facing the US today is bad data. Anyone relying on initial government reports literally has no idea what’s is really happening.

And here’s why. The two most important numbers on the US economy are Gross Domestic Product, put out quarterly by the Bureau of Economic Analysis, and the monthly jobs and employment report from the Bureau of Labor Statistics.

I will try to keep this simple. Some of you know that about 70 percent of GDP comes from determining what the BEA calls Personal Consumption Expenditures. PCE is most everything other than capital goods that we spend money on during each quarter. But what almost none of you know is that the BEA gets its spending data from the Census Bureau’s monthly survey of all of 12,000 retailers out of more than three million retailers. That is about 1/3 of one percent of all retailers.

I actually emailed the BEA and asked instead of surveying 12,000 retailers why not use Master Card, Visa, AMEX and the like which has over 95% of actually monthly retail sales? What’s more Master Card and Visa actually sell aggregate sales data.

I was surprised to get an email answer, full of bureaucratic reasons why change is difficult, from the Associate Director for National Economic Accounts at the BEA. The eamil ended by assuring me that discussion about whether real time data vs surveys has been and is going on. Yes, I will bet those discussions will continue for years; or until someone tells them to just use real time data and stop the survey nonsense.

Similarly, the BLS monthly employment report is meaningless, full of sound and fury, signifying nothing. The unemployment percentage is derived from the Census Bureau which surveys a mere 60,000 households each month to determine the number of people working, in the workforce, etc. Sixty thousand households are surveyed. That is it.

Oh, yes. The BLS also surveys about 140,000 business and government entities every month and gets about 100,000 responses the first month to guess at how many new jobs are created. And to give the BLS some credit, in the footnotes to the monthly report, the BLS does admit that the there is a 90 percent chance that the actual jobs number was within 100,000 of the reported number.

That 90 percent chance of being accurate within 100,000 jobs is not very accurate at all. And yet the media and Wall Street treats the BLS number as if it means something.

There’s a much better way of doing this. Remember, every pay period all employers send in to the US treasury withheld income and employment taxes. If all employers included with the tax payment the number of people working and how much they made, we would have wage and salary and jobs data in real time. We would not need surveys, nor government economists more concerned with their jobs and the status quo.

All real progress starts with accurately representing what is going on now. Without really knowing what is so, all we have is the confusion, uncertainty and the mess existing today.