Everything's Fine... If You Ignore History

Submitted by Salil Mehta via Statistical Ideas blog,

This must be the "new normal" everyone?'s talking about?. Where a cyclical low ?in the unemployment rate?, to the low 5 percent range?,? is met with a courtly high stock market and talks of an impending rise in rates. Sure there is a feeling in some pockets ?of the economy ?that "things" are much better out there, even though by some measures the ?labor data is unfortunately, still stuck in a rut even through the global financial crisis.

?For example, ?Q1 U.S. GDP contracted? again? (and by nearly 1 percentage point more than GDPNow had originally estimated!) And despite Wall Street cheer, Q2 GDP ?may only come in at about 1 percent.  ?And on the tangible employment front, it is important to look at alternative measures of employment to get a proper read on how things genuinely align to a time when we all shared similar "good feelings" in our past.  ?After all, a feeling of "better" should often be associated with a significant proxy in the past that serves as a benchmark for that term.

?Let's look at the employment to population ratio. Th?e? ratio ?for May just ticked up 0.1 percentage points, to a cyclical high of 59.4 percent. Still, this level is well shy of the 62.0 percent? that was the? worst reading reached ?in? 2003 (in the aftermath of the 2001 recession). Don't be fooled by this conservative estimate of being only a 2.6 percent point difference; on a labor force of 157 million that is equal to 4 million not-employed people.


The official unemployment rate doesn't take into account the broader population, so even though ?its current 5.5 percent ?appears to be an improvement (the level resides between the previous cycle's high of 6.3 percent in 2003 and ?its?? low of 4.4 percent ?in? ?2006-? 2007), we must seek a broader unemployment measure to better reflect the economy's changing labor force (or should we say the chang?e in who's not in the labor force?)

The U-6, the official broadest formulae for unemployment and underemployment in the U.S., is at 10.8 percent.  ?This? level is still completely higher than the worst reading we saw in 2003 (10.4 percent). In fact prior to the current cycle, we have to ?go back in time 2 cycles (a lengthy 21 years) in order to see a U-6 reading that is ?even worse than ?the current ? 10.8%. Given the inclusion now, of the marginally attached people from the labor force, every 1 percent difference in the U-6 rate is also not a small amount; it is equal to 1-2 million Americans negatively impacted (and we are still a couple percentage points higher on U-6 versus where we'd prefer to be.)


What this all evidences is that in some elements of the economy, and some influential portions of the population, things feel great (perhaps as good as ever!) But for a few million people in ?the economy it still feels depressing.  Just looking at the fairly, well appreciated U-6, and seeing that is now finally within a 1/2 percentage point (of the worst ?from the 2001 ?recession? fallout) and thinking some are forced to embrace th?is now as a sign of a recovered market. Something seems premature when celebrating the employment to population ratio, or even the U-6, both of which are just as easy to see alongside the headline payroll numbers. It's because at the end of the day, the "new normal" shouldn't be one where our current best is still worse, then our prior worst.