Let’s get one thing straight. I am no economist. If I were, I’d be too embarrassed to tell anybody anyway. I’m just a guy who learned from Yogi Berra that, “You can observe a lot by watching.”
I do like to watch, but not financial “news” TV. It’s a waste of time and counterproductive, especially CNBC infomercial TV. I became fearful that if I watched too much of that stuff I would grow hair on my palms and drool green sputum. So I turned off the TV 10 years ago. I do admit to listening to Bloomberg radio from time to time, however.
The thing I like to watch is data. It sounds boring, but it just depends on how you look at it. I like the raw, unedited material, suitable for mature audiences only. None of that seasonally adjusted cartoon stuff, no sir! I like the hard stuff. The rawer and fresher the better. The sharper and more jagged the edges, the more it excites me.
To make matters worse, horror of horrors, I like to compare streams of raw unedited data from entirely different and completely unrelated things, like say, the Fed’s balance sheet, the stock market and the economy. All on one chart! Such dirty stuff. I make sure the door is locked and the shades are drawn. But somebody has to do it, and I appointed myself. Yes, it’s a perversion. But I just can’t help it. I look at these twisted lines for hours on end. And I see stuff.
One data series really excites me because it comes out daily and it relates to the real economy. It shows how much the government is collecting every day in withholding taxes. That’s right! Every freaking day the government reports how much it takes in in withholding and other taxes. It also tells how much it spends by category, daily.
Now, who else has told you that? CNBC? No. Bloomberg? No. Dow Jones? Pfffft. Of course not! How about Reuters? They’re upstanding people in the community. But no, no reports there either. In fact, no mainstream media types, and to my knowledge, no major brokerage firms or banks report this data. They have it. The US Department of the Treasury publishes it every day, but the only person in the Wall Street mainstream whom I’ve seen discuss it is Barry Ritholtz. He gets it from Matt Trivosonno who’s the only other person I know of who has been covering it. That’s a pretty small universe of coverage for something so important. [Ed. note- I was subsequently reminded that Charles Biderman also covers this data. Apologies to Mr. Biderman. Still a small universe.]
Not only that, but it’s clear that even if Wall Street’s economic talking heads know it exists, and have it, they pay no attention to it whatsoever. There’s just no other way to explain how the Street punditocracy could so consistently underestimate the strength of the economy for most of the past 6 months. It seems like every week the economists’ are too pessimistic. Economics is alchemy and economists are charlatans, quacks, frauds, and shills, but still, you would think that they could get lucky and get it right once in a while. Had they been paying attention to this readily available data, there’s almost no way they could have gotten it wrong. But they did and they’ve continued to.
I look at this data on a daily chart with a couple of moving averages to quiet things down because the data is so choppy. It’s choppy, but it’s not noise. The patterns are repetitive and follow a highly esoteric and inexplicable rhythm which you have probably never experienced. These patterns have the technical name “pay periods” and they occur either once a week, every two weeks, twice a month, or once a month. I know it’s technical, but stay with me here. It means that collections are higher on Fridays than on Mondays! They are higher on the 15th and 30th or 31st of the month, than other days. Do you realize how how much research was involved in discovering this?
Aha! I didn’t think so.
So yes, the data is choppy, but it follows regular patterns, and if you line up this year against last year and smooth out the peaks and troughs with just a little moving average smoothing it’s not that difficult to get a reasonably clear picture of the trend. Of course, horror of horrors, 365 isn’t evenly divisible by 7, the number of days in a week, so that means that every so often it’s necessary to time shift the data a wee bit. Is that so tough? Trust me. It isn’t.
Given that background here’s what the data looks like as of now. I start with a two week moving average, based on what I recall as a typical “pay period” from my working days. As of Friday, March 22, this year was 10.6% higher than last year. I then smoothed those series to a one month moving average. As of Friday that line was 11% above last year. Very impressive gains, I must say.

Of course some of that is due to the increased tax rates that went into effect on January 1. Rather than bore myself with an analysis and discussion of the accounting on that, I thought that a little before and after picture would work. There was some fiscal cliff beat the clock tax stuff going on in December, which was offset in late January, so I looked at November 15 as the before and February 15 as the after. There was a jump of 6.5% between those two dates. Given the little that I know about the tax rate changes, and doing the arithmetic in my head, that seemed about right.
If we assume that 6.5% is the amount of the gain attributable to the tax increase and deduct it from the current 11% margin over last year’s March 22 collections, it suggests that the remaining 4.5% is due to a combination of wage and salary inflation and economic expansion.
Getting a handle on inflation when it comes to wages and salaries is a little tricky. I like to use the BLS data on average weekly employee compensation because it’s the most current data. It rose 2.1% in February. That’s fairly close to the 12 month average rate of increase. So if inflation accounted for 2.1% of the increase in total taxes withheld, that would mean that economic expansion would account for the difference of 4.5% minus 2.1%, which equals 2.4%.
So there it is. The economy is growing at 2.4%. The devil is in that wage and salary inflation figure though. Allowing for a little fudge one way or the other, the economy could be growing at 2% or closer to 3%. Estimates below that range would seem too pessimistic. And nobody’s higher than that range. Virtually no one that I’ve heard of is irrationally exuberant about the economy, although that in itself is probably an indication that we’re headed that way. Those people do have a tendency to be wrong about everything, and when they’re all too low with their guesses, it suggests that in reality things are getting hotter than they think.
Here’s what the difference between this year’s withholding and last year’s looks like as adjusted for inflation. You can clearly see where the 4 week moving average jumped from around +5% in December to over 11% in January from the tax increase. After some payback in early February for the fiscal cliff tax increase front running that took place in December, the numbers have bounced back and remained pretty stable near 9-10% since late February.

It’s funny that economists try to guess these numbers to within a tenth of a percent. The media pushes them with surveys on every economic data series under the sun. It’s a little arrogant, and mostly just crazy to pretend that anyone could guess these numbers to that degree with any regularity. Crazier still is that we pay attention to all the silliness and try to trade it.
Please add your comment below.
This post is adapted from the weekly Wall Street Examiner Professional Edition Treasury Update, available to try for 30 days, risk free.
Stay up to date with the machinations of the Fed, Treasury, Primary Dealers and foreign central banks in the US market, along with regular updates of the US housing market, in the Fed Report in the Professional Edition, Money Liquidity, and Real Estate Package. Try it risk free for 30 days. Don’t miss another day. Get the research and analysis you need to understand these critical forces. Be prepared. Stay ahead of the herd. Click this link and begin your risk free trial NOW!
Follow my comments on the markets and economy in real time @Lee_Adler on Twitter!
Lee, your on top of your game with this one, very creatively written.
Thanks NE! You get me.
I try to have fun with this stuff, and hopefully some folks will have fun reading it. Unfortunately, however, some people have no sense of humor.
Charles Biderman of Trimtabs has used withholding taxes as a tool for judging the economy’s performance for many years. In fact, he was the very first person to do so to my knowledge. Just to set the record straight with regard to ‘nobody is doing that’.
You’re right. I forgot about him. I tend to discount him because I recall him being wildly bullish at the top in 2007. I have a bad habit of discounting people who have been spectacularly and adamantly wrong for all the wrong reasons.
Biderman was strictly an institutional guy until a couple years ago, getting $30k per year for his service. I never saw him go public with this until about a year ago when I first heard his economist on Kathleen Hays’s show.
As far as him being the first, I know that Trivossono has been doing it for years and I started tracking it for my subscribers about 4 years ago.
I think Liscio Reports also covers it.
Outside of Kathleen Hays interviewing Biderman’s economist I’ve never seen this discussed in the mainstream media. I apologize for taking a bit of literary license and not mentioning everyone who has ever covered this subject. Indeed Mr. Biderman does.
test
I like the analysis, i’ m not sure that a payroll increase directly correlates to economic growth, although it is certainly a factor.
Thanks. Withholding taxes don’t every absolutely everything, but they cover, or at least represent, most economic activities, so they’re a pretty good indicator, except when skewed by policy changes or other external factors. There is some withholding, usually a non material amount, that is from sources other than payrolls that occasionally can be higher than normal background levels, which makes it tricky.
Lee, I’m not sure I buy into this analysis for several reasons.
Your analysis suggests that increased tax revenues is indicative of increased GDP, specifically the difference between tax receipts year-to-date in 2013 and year-to-date tax receipts in 2012. Your analysis makes the assumption that 6.5% of the increase is due to tax increases and 2.1% of the increase is due to inflation, leaving 2.4% the result of GDP increase. These are big assumptions.
It is quite likely that the magnitude of tax increase is much larger than 6.5%. We know that Uncle Obama clipped every paycheck with a 2% increase in payroll withholding and an additional 0.9% medicare tax on those earning more than $200,000 (which account for over 50% of all federal payroll tax receipts). In addition the taxes paid on dividends and capital gains skyrocketed.
Historically, tax receipts make a spectacular jump on January 15 because quarterly filings are due on that date. People who liquidated holdings in late 2012 to avoid the steep capital gains increases in 2013 would likely file quarterly returns on January 15, as would those individuals who dipped into their 401K accounts to make ends meet in this very bad economy. The size of the tax and penalty payments for these transactions are yet unknown, but they could prove to be substantial.
Another unknown is the effect of greater IRS compliance audits of the Earned Income Tax Credit program, which is arguably the most fraud-ridden aspect of the tax code. Increasingly the IRS is doing its job and rejecting tax filings that make dubious claims for EITC credits. This has the net effect of raising the tax bills of individual Americans.
The 2.1% inflation assumption is laughable.
A 2.4% GDP increase on the basis of increased tax revenues? I don’t know. All this article can state for certain is that the government raised taxes and is collecting more money.
The analysis is about withholding taxes, which reflect real time employee compensation in the US economy. Withholding taxes do not include quarterly filings. Quarterly estimated tax filings have no impact on them. The EITC has no impact on them. They’re not relevant to the analysis and the point of the article.
Also, your math is fuzzy on the tax increase. The actual increase post tax increase is pretty clear, and it squares with the math.
Where you dispute that wages and salaries have risen by some amount other than the 2.1% BLS figure for year to year change in February, you left that hanging. Do you think that compensation grew more or less than that? If less, then growth would be more than 2.4%. If compensation inflation was more, then econ growth would be less than 2.4%.
While my headline implied GDP growth, the article really wasn’t about that. It was about the fact that Wall Street conomists have been consistently underestimating all economic data series for the past six months or so. But wouldn’t it be interesting if Q1 GDP growth did come in at 2.4% or something close. Then I’d say I did a good job. If it’s way off, then not so much. And let’s wait till the final GDP is in before judging.. The first estimate is always way off.
What about the “more than 1 million jobs created in 2112”? How do they factor into the calculation? Still represent an expansion. But does the 2.4% growth math still work?