Something new here too. Since 2006 wage gains among production workers are stronger than for all workers in all but four sectors, which suggests, of course, that managers are falling behind.

The Bureau of Labor Statistics has wage data for manufacturing production workers running back to 1939. In 1939, average hourly earnings for nondurable manufacturing were just 52% of those for durable manufacturing, 34 vs. 65 cents, but rose almost straight up in the 1940s to 80% and continued to drift up to their current 89%.

In the Trade, Transportation and Utilities sector, however, retail wages have lost ground, falling from 84% of TTU wages, to 80% currently, while wages for transportation & warehousing have fallen from 127% to 110%. Wages for wholesale trade have risen from 109% of TTU wages to 126%, while utility workers have seen their wages climb from 132% to 184%.

Back in 1964, when records were established for the goods producing and service providing sectors, wages were even, but by the 1980s wages for services had slipped to 80% of goods. They began a rocky climb back up and have levelled out at about 94% currently. Since 1964, construction and manufacturing wages have risen by similar percentages, but logging & mining wages have gained an additional fifteen percentage points.

We’re looking back to 2006 now because that is as far back as the all-worker series goes. Since then, wages in logging and mining have made the strongest gains, up 49%, to $28.96, followed by leisure & hospitality, up 47% to $14.11. Production workers in finance and in professional & business services made similar percentage gains, to about $27.00, as did construction work, to about $28.00. Since 2006, wages in services are up 40% and in goods up 36%.

Weakest gains for production workers are in manufacturing, transportation & warehousing, and retail trade, all up around 30% to $21.77, $22.07, and $16.15 for retail.

Since 2006, however, production workers have seen greater wage gains than all workers in all but four sectors, so managers are dragging that broad tranche down. In manufacturing production workers lagged all workers by 1 percentage point, in durable manufacturing by 3pps, in utilities by 5pps, and in information by 11pps. In leisure & hospitality, production workers were 6pps ahead; in retail trade, 5pps; and in logging & mining, construction, and education & health, all ahead by 4pps.

Over the last year, production workers wages are up over 5% in logging & mining, up 4% or more in information, retail trade, leisure & hospitality, and construction. In only other services, 1.9%, and finance, 1.6%, are they up less than 2 percent.

Over the last year wages for all workers in information, a struggling sector by other measures, are up 6%, up 4% or more in construction, retail trade, leisure & hospitality, and finance, followed by a string of 2-3% gains, with non-durable manufacturing, and transportation & warehousing reporting the weakness growth at 1%.

The wage picture is certainly complicated, with some of the largest percent gains occurring side by side between the highest and the lowest hourly wages, and quite a few “not as we expected” trends. By a number of metrics the different sectors are going in different directions, which further confuses an already cloudy picture.

First, party: The University of Michigan collects data by political party only sporadically, so we don’t have a full history but, as we brought up at the time, last May, Richard Curtin, the guy who puts it all together, noted that in fifty years, the survey had never recorded as “dominant a political effect,” as it did in early 2017. Curtin expected that divergence to converge, but instead it widened.

Here’s the graph we ran:

During the years Trump has been President, Republican sentiment has averaged 117.6; while that of Democrats averaged 80, meaning Republican sentiment is running 46 points above the Obama years, while sentiment among Democrats is running just 13 points below. Both calculations exclude the depth of the recession years, when the divergence straitened to just 9 points; whoever thought we’d remember anything reassuring about those years.

That effect continues. In an October presentation Dr. Curtin identified his “major underlying issue,” as whether economic expectations surveys can retain their predictive ability given their “responsiveness to political rather than economic developments.” He quotes some observers who believe the partisan effect is “uniquely tied to the Trump administrations,” but his research suggests it is tied to our old friends income inequality and wage stagnation.

Second, educational attainment: Curtin also revisited the fact that the partisan divide between those with college-degrees remains “very low and insignificant,” and built on the striking observation he made back in May: the demographics really changed in 2017. Through 2016, generally, assessments of economic possibilities rose with income and education, and fell with age.

That changed in 2017.

Overall expectations for those with less than a high school degree had risen from 68 in October 2016 to 82 by August 2017, but fell for those with a college degree from 87 to 80. For those 65+ they rose from 2016’s 69 to 83 in 2017, while slipping a bit, to 86, for those 18 to 34 years old.
May 14, 2018
Sentiment for all three education terciles tracked each other closely throughout the series, all peaking around 2000, high school at 100 and the two college groups at 120, before falling raggedly to about 60 in the recession. The current divergence is led by an increase in the outlooks of the two lower attainment levels while those with a college degree are basically flat.

Curtin suggests that those with “relatively low job skills, as proxied by education, were the most affected by Trump’s election.” As we mentioned when we first brought this up, it’s a good thing if people with lower skills can do better.

But it’s a very bad thing if partisanship is putting a dent in the value of formerly trusted economic data.

First, party: The University of Michigan collects data by political party only sporadically, so we don’t have a full history but, as we brought up at the time, last May, Richard Curtin, the guy who puts it all together, noted that in fifty years, the survey had never recorded as “dominant a political effect,” as it did in early 2017. Curtin expected that divergence to converge, but instead it widened.

Here’s the graph we ran:

During the years Trump has been President, Republican sentiment has averaged 117.6; while that of Democrats averaged 80, meaning Republican sentiment is running 46 points above the Obama years, while sentiment among Democrats is running just 13 points below. Both calculations exclude the depth of the recession years, when the divergence straitened to just 9 points; whoever thought we’d remember anything reassuring about those years.

That effect continues. In an October presentation Dr. Curtin identified his “major underlying issue,” as whether economic expectations surveys can retain their predictive ability given their “responsiveness to political rather than economic developments.” He quotes some observers who believe the partisan effect is “uniquely tied to the Trump administrations,” but his research suggests it is tied to our old friends income inequality and wage stagnation.

Second, educational attainment: Curtin also revisited the fact that the partisan divide between those with college-degrees remains “very low and insignificant,” and built on the striking observation he made back in May: the demographics really changed in 2017. Through 2016, generally, assessments of economic possibilities rose with income and education, and fell with age.

That changed in 2017.

Overall expectations for those with less than a high school degree had risen from 68 in October 2016 to 82 by August 2017, but fell for those with a college degree from 87 to 80. For those 65+ they rose from 2016’s 69 to 83 in 2017, while slipping a bit, to 86, for those 18 to 34 years old.

Sentiment for all three education terciles tracked each other closely throughout the series, all peaking around 2000, high school at 100 and the two college groups at 120, before falling raggedly to about 60 in the recession. The current divergence is led by an increase in the outlooks of the two lower attainment levels while those with a college degree are basically flat.

Curtin suggests that those with “relatively low job skills, as proxied by education, were the most affected by Trump’s election.” As we mentioned when we first brought this up, it’s a good thing if people with lower skills can do better.

But it’s a very bad thing if partisanship is putting a dent in the value of formerly trusted economic data.