Just Two Recession Indicators: Shamelessly Spinning the Data

I received an email from a reader asking my opinion of a Zero Hedge article published yesterday: Just Two 'Recession' Indicators, the two being retail sales and personal income less transfer payments. The former takes its message from an eight-year chart of retail sales for clothing and general merchandise lifted from BloombergBriefs.com (which requires a $1495 subscription, so it's not on my daily reading list). The second highlights a widely circulated chart from David Rosenberg.

I routinely track real retail sales in a couple of monthly updates, the first of which is also in conjunction with real personal incomes less transfer payments:

  • The Big Four Economic Indicators
  • Retail Sales: The "Real" Story
  • And I also drill down further on personal incomes:

  • Disposable Income Per Capita
  • Please note what Zero Hedge's eponymous Tyler Durden has to say about real personal incomes less transfer payments:

    [A]s Gluskin Sheff's David Rosenberg points out real personal income net of transfer receipts plunged at a stunning 5.8% annual rate in Q1. The other seven times we have seen such a collapse, the economy was either in recession of just coming out of one [source].

    Scary, huh? Especially for those readers who fail to recognize a bit of shameless sensationalism. Here is an overlay of the Big Four economic indicators that I regularly follow showing the monthly percent change since the end of the Great Recession. Let's focus on the red line: Real Personal Incomes less Transfer Payments.

    I added a linear regression to highlight the erratic movement of the PI less TP series. As I've repeatedly explained elsewhere, the big surge in incomes in November and December of last year was the result of incomes being drawn forward in Q4 as a tax planning strategy in expectation of higher taxes (the Fiscal Cliff).

    Here is my own replica of the Gluskin Sheff chart, using the same Annualized Quarter-over-Quarter Percent Change manipulation of the data.

    First of all, a QoQ percent change completely obscures any traces of the tax strategy that is clearly reflected in the monthly data. Secondly, David Rosenberg's chart annualizes the QoQ percent change, which is the normal (if somewhat strange) manipulation we expect for quarterly GDP, but not for routine monthly economic data series. And what does that quarterly annualization do? It gives us a BIG negative number: -5.8%.

    What does the chart look like if we do a simple and more intuitive QoQ percent change without annualizing it? Have a look. The numbers for the outlier quarters shrink substantially: That -5.8% becomes a far less shocking -1.5%.

    Since a decade would pass before the next recession following the Q1 1993 "plunge", it would presumably fall under Tyler Durden's "just coming out of a recession" category (of course, that would be a whopping 21 months after the eight-month recession that ended in March 1991, but time is relative, no?).

    What was the real cause of the Q1 1993 "plunge" in real PI less TP? The same sort of tax management strategy that explains the Q1 2013 drop in the annualized quarter-over-quarter percent change manipulation of the data. Businesses were motivated to pay bonuses before the end of the year and pull all sorts of earnings forward to avoid the tax increase. Here is a typical article from the times (the LA Times, to be precise) published during the last week of 1992: A Tax Cloud Hangs Over the 1992 Bonus.

    What was the historical context? In 1992 the top tax bracket for married filing jointly was 31% starting at $86,500. In 1993 two higher tax brackets were added: 36% on $140,00 to $250,00 and 39% on $250,000 and higher.

    Let me close this commentary on a personal note. I'm deeply concerned about the trends real income. It's a topic I study in several contexts:

    But to focus on the annualized quarter-over-quarter percent change in PI less TP to make an ominous point about Q1 2013 is, as I stated earlier, shameless sensationalism.