New economic data only deepens the mystery.

Summary: When will the next recession arrive? Economists, like most professionals, sell certainty. They assure us that data has meaning and can be understood. Sadly that’s reliably true only during trends, so economists usually miss inflection points. And a new era began ~2000, bringing new dynamics and normals, making the search for inflection points especially difficult. Bottom line: we have more questions than answers. Embrace the mystery!  This post will be updated.



  1. Recession watch?
  2. Wholesale Sales for January
  3. Retail Sales for February
  4. Mfg & Trade sales & inventory for January
  5. What do prices might tell us?
  6. Conclusion
  7. For More Information

(1)  Recession Watch?

This expansion has run for 68 months, the 5th longest of the 11 recessions since WWII (average length of 58 months). It’s time to start our recession watch. The next downturn might be exciting, given the weakness of the economy and that we’re starting it with rates already at zero.

The GDPnow model of the Atlanta Fed forecasts Q1 GDP at 0.6% and falling fast:

Atlanta Fed GDPnow forecast

(2)  Wholesale Sales for January.

The recent data consists of alternating good and bad news (see the posts at the end for previous posts about this). Optimists obsess about the strong jobs number, but that’s only one metric — and a coincident indicator (unemployment claims is a leading indicator, and running flat).


First up today we look at a dark dot: wholesale sales for January, down 3.1% MoM SA. Despite the easy excuse of it’s just oil, the decline affected both durable and non-durable goods (a broader range than the slowing in December). January’s 1.4% drop in durable goods MoM SA was the 3rd largest since the recession (behind Feb & May 2011).

As usual with these unexpected downturns, inventories accumulate. See the ugly chart below of the inventory/sales ratio, showing the kind of numbers that excite CFOs (cut production! push sales!). Petroleum is not driving that rise: its i/s ratio is up 25% YoY, but it is only 3% of total inventories and 10% of sales. Remember the context: this is January data (February was probably worse), and the i/s ratio is a lagging indicator (although not part of most indexes).

Wholesale sales/inventory ratio: January 2015

(3)  Update: Retail Sales for February

February retail sales were -0.6% SA MoM, following December’s -0.9% and January’s -0.8%. Feb was different,however: gas station sales were not a factor in February (they were up following their crash in Dec and Jan). Retail sales are a coincident indicator.

See Alhambra Partners for a detailed analysis of these ugly numbers. January and February YoY retails sales ex-autos NSA were the worst since 1992 other than those months affected by the Great Recession — and September 2001 (the tech bust and 9/11).

February 2015 Retail Sales

(4)  Update: Manufacturing & Trade sales & inventory for January

The report was ugly across the board, confirming the data from wholesalers shown above. January was down both MoM (-2.0%) and YoY  (-0.3%. So far few numbers have gone red on a YoY basis. The inventory/sales ratio is rising fast; it’s a lagging indicator.

January Manufacturing and Trade investory to sales ratio.

(5)  What do prices might tell us?

I don’t share the avant garde fetish for market signals, since they’re so often noise when they’re not mysterious. But they’ve got to be watched.

Interest rates should be one of our best guides, but they’re hopelessly distorted by the machinations of the world’s major central banks. Even credit spreads tell us little but that CB’s zero interest rate policies (ZIRP) — for some CB’s, now a negative rates policy — have forced even risk-adverse investors into the deep end of the pool. As Guy Haselmann (capital markets strategist at Scotiabank) says: “Unlike in years past, neither central bankers nor market participants can extract any information from current bond valuations.”.

Watching stock prices has been a waste of time for years as they move with the Fed’s monetary magic, and that’s been “on” since the depths of the recession. Someday they will again become useful. We watch stock prices hoping to see that day before events run over us (much as Wile E. Coyote peers into the dark railroad tunnel into which the Road Runner just escaped).

The S&P 500 index has returned to the floor of its trend ( the green line, “support” if you believe such terms have meaning). Watch for a breakdown. US stocks have had the 3rd greatest 6-year run since 1900, behind only the periods ending in 1929 and 1999. That doesn’t mean this will end badly, but suggests keeping our eyes open.

S&P 500 price chart
From, posted with permission.

Other prices are similarly opaque. The global trade numbers look strong (through December, at least), but the Baltic Dry Index is weak. It shows the cost of moving bulk cargo by ship (e.g., coal, iron ore). It was 509 on February 18, the lowest since the record began in 1985.

Industrial commodity prices are weaker than they should be in a booming global economy.  Look at copper (aka Dr. Copper, an honorary degree for its skill at economic forecasting). The scale shows the price gain since the recessionary trough. Copper rose through 2009 and 2010, peaked in 2011, and has since lost 60% of its gains.

Spot copper prices
Spot copper prices from; with permission.

(6)  Conclusions

I don’t know what’s happening, and I doubt than anyone does. That’s the tell, the vital message of markets to us since the crash. We’re in a new age, which brings new normals. Despite the hoary wisdom, sometimes it really is different this time. Regime change runs slowly until it becomes rapid, then instantaneous. Nobody rings a bell to announce it.

(7)  For More Information.

(a)  See all posts about economics. Especially see this: Dreams of a boom fade & attention turns to secular stagnation. Also see these posts about “stall speed”: The dilemma of the US economy: can’t take off & too close to the brink and Has America’s economy entered the “coffin corner”?

(b)  Recent posts about this cycle:

  1. Dreams of a boom fade & attention turns to secular stagnation.,
  2. Status report on the US economy: darkening sky, rough waves ahead.
  3. How close are we to the next recession? — More indicators.
  4. Highlights of the jobs report, the good news & the bad.



11 thoughts on “New economic data only deepens the mystery.

  1. You are quite right, which is why comparisons with 1950-2000 ‘postwar’ recessions won’t work. This is a Depression albeit not as severe as that of the 1930s and 1890s.

  2. One possible definition of “Depression” is an economic conjuncture that requires significant structural and institutional changes before normal growth resumes. That is, not just an inventory buildup or temporary shock. As you have pointed out this time there are all sorts of new phenomena this time.

    1. Social,

      Yes, there’s no standard definition of “depression.” But I suspect you’d have to look very hard to find many past events called “depressions” by an economist that meet your definition with the ~2.5% GDP growth the US has had since the trough of the downturn. My guess is that you can not find any.

      While below the post-WWII average, I think it is not remotely appropriate to call 2.5% growth a “depression”, especially when most metrics have exceeded the pre-crash peak. Using that word certainly gives a misleading impression to non-economists who hear it — which is imo the goal of most people using it.

    2. “I suspect you’d have to look very hard to find many past events called “depressions” by an economist that meet your definition with the ~2.5% GDP growth”

      There is one: the Long Depression, 1873-1896. During that period, the real US GDP grew on average 3.52% per year. In France, Germany and the UK, the growth rates were 1.73%, 2.46% and 1.93% respectively.

      Things become much more interesting when looking at per-capita GDP growth, which over the period was only 1.3% for the USA. We can further distinguish three periods: an initial low 1873-1878 at 0.37% yearly growth; a brief recovery 1878-1880 at 8.77%; and a long slump 1880-1896 at 0.59%.

      All figures taken from a historical dataset from the Groningen Growth and Development Centre (large spreadsheet). It would be interesting to look at the aggregate and per-capita GDP growth rates since 2007 for a comparison.

    3. guest,

      I had heard about the Maddison Historical Statistics project, but never looked at it. Thanks for the link!

      Color me skeptical about reconstructions of GDP for the 19th century. We have almost no economic data — as we consider it today — from that period. We have spotty data on some large manufacturers, ag output and prices, tariff data — and the decennial census. But calculations of a high level number like GDP should be regarded skeptically — esp when comparing it to current data.

      Giving the GDP changes in 19th C GDP to 3 significant digits is absurd. The Census of Manufactures began to collect data on hours worked and wages in 1880. Reliable factory output indexes start (from memory) around WWI. Reliable surveys of unemployment numbers start aprox WWII (note the pattern here). The project to create the great National Income and Production Account tables began in the late 1920s, was made a priority in WWII, and the first formal national accounts numbers were published by the US govt in 1947.

      I’d like to see an actual historian comment on this before drawing conclusions, not just the output of a database.

    4. guest.

      “there is one: the Long Depression, 1873-1896.”

      Here’s your mistake. The Long Depression is defined as 1873 – 1879, not 1896. See the official dates from the NBER. see Wikipedia).

      See this description of it by economist Timothy Taylor:

      The US economy was in a continuous recession for 65 months from October 1873 to March 1879. Historians call is the “Long Depression,” because the Great Depression from 1929 to 1933 saw “only” 42 consecutive months of economic decline. For comparison, the more recent Great Recession lasted 18 months.

      Samuel Bernstein offered one of the classic descriptions of the Long Depression in his 1956 article, “American Labor in the Long Depression, 1873-1878” (Science & Society, Winter 1956, 20:1, pp. 59-83, available through JSTOR). Precise government statistics are not available for this time period, of course, but estimates of the unemployment rate for the later part of this period often exceeded 20%, and some exceeded 30%. For those with a job, real wages fell by half. Even those real wages were often paid in the form of company scrip, which could only be used at the company store, and was worth substantially less than cash.

Leave a Reply