January Seasonal Blues: Employment Dropped -2.84 Million In January, Construction Dropped -272k

Confounded Interest ^ | 02/03/2013 | Anthony B. Sanders

[whitedog57]

Every year, employment drops by almost 3 million during January. While the drop was worse than 2012, it is on par with the drops in 2010 and 2011.

Of course, the Census Bureau whips out the old X-12 ARIMA smoothing models (maybe even the new and improved X-13!) and produces this much more palatable chart.

Phew! That is much better for the media (and economy cheerleaders like Mark Zandi) to tout. Thanks to the ARIMA smoothing model, employment INCREASED by +157.000!

As my students know, I like the raw, unsmoothed numbers. All the smoothed numbers tell you is … over a long time period, this January was better than previous January employment numbers. But it masks the fact that employment actually fell by -2.84 million in January 2013.

Construction

The same story holds for construction. According to the ARIMA model used by the Census Bureau, the US gained 28,000 employment jobs in January. But the raw data indicates otherwise: construction employment dropped by -272,000 in January.

Yes, this is the classic construction seasonal. Construction slows down in the winter (particularly in the frozen north) but generally heats up in the Spring and Summer. So once again, the ARIMA model produces a much more palatable representation of the construction numbers.

In terms of construction employment, the US is at February 1997 levels. This is consistent with other housing indices: we are at pre-bubble levels.

So while Seasonally Adjusted employment numbers (ARIMA) look nice for the media and pundits, the raw numbers are the real deal.

But the policy point is that The Fed and Federal government should avoid create massive construction and housing bubbles, if they can. And avoid public policy like the deadly National Homeownership Strategy. nhsdream2

[whitedog57]

Sheesh! But great charts!!!!