You go, BLS!
The Bureau of Labor Statistics just released a new data series derived from the Employer Costs for Employee Compensation series. These ECEC data have been around for awhile–they’re the basis for the more commonly cited Employer Cost Index data*–but only for averages (see Technical Note here if you want more info). Now we have compensation, wages, and benefits for the 10th, 50th (median), and 90th percentile. I consider these to be high quality data that, relative to other series, provide a more fulsome look at the full compensation package.
At this point, the data only go back to 2009, with first quarter data through this year (i.e., all the data that follow are for the first quarter of the year in question).
And now, without further ado, here’s my initial analysis of these data.
First, we have real comp, wages, and benefits for the three available percentiles, indexed to 100 in 2009.
The real comp figure shows low pay falling faster than middle or high pay as the recovery got underway. Comp starting growing around 2014, but faster for high-end (90th percentile workers). As we’ll see, high-end wages and especially benefits outpaced the other groups, implying, as shown in the last figure, higher wage inequality. As other data have shown, such as the monthly wage data for production, non-managerial workers, real median comp has grown slowly since 2016.
Next, the real median wage had but one year of strong wage growth–2015, when inflation was close to zero–and is now just back to where it was in 2009. Both low and high wages did better, up about 5 percent over the period. Still, that’s about 0.5 percent per year, well below productivity growth for this period (1.2 percent per year). As I’ll show in a forthcoming piece–an important finding, I’d argue (and one Elise Gould of EPI found first, ftr)–recent gains in low wages occurred mostly in states with higher minimum wages.
Benefit growth (and benefits shares, as I’ll show in a moment) reveal particular disparity over this period, up over 15 percent for high-end workers, 5 percent for the median, and still below 2009 levels for low-wage workers.
For both middle and high wage workers, benefits account for about a third of compensation, but for low-wage workers, it’s just a fifth. Benefit shares grew slightly for middle and high-wage workers from 2009 through around 2014, but contrary to claims by Trump’s CEA, benefit shares haven’t gone up at all over the past couple of years for any wage group.
This information, btw, on benefits by wage levels is both important, new, and, while these findings support many of our priors, they’re revealing of the point that when you’re talking about compensation, including benefits, you can’t just blithely cite averages and think you’re saying anything about outcomes for low-wage workers.
This last figure shows how compensation inequality has grown over this period–a continuation of a multi-decade trend–as high-end pay has outpaced that of middle-wage workers (it’s just the ratio of 90th percentile comp to that of the median).
Finally, here’s a data table with levels in real 2018Q1 dollars, along with annualized changes.
More to come as I and others have time to dig into the guts of the data. EG, the BLS provides breakdown of benefits by type for each wage group, not to mention, industry and occupational level data.
Forgive me if I’m inordinately excited about this release, but in a era where facts are on the run in this benighted town, these new data provide a welcome reminder that there are still some folks around here who are quietly and effectively doing their jobs.
*The ECI data are the ECEC data with fixed weights for occ and inds, so as to capture a “purer” measure of comp changes that are not due to ind and occ upgrading.