Thinking about the Fed, Wages, Part 3

Fed Focus

 Those urging the Federal Reserve to delay include those who believe the economy is too fragile to withstand any increase in rates, and that our more vulnerable workers will take the hit, and those who believe the stock-market will be in for a fall. 

 We’ve argued the weak wage case repeatedly, and will spare you a recap. Some of our most respected analysts argue that wage increases are a late-cycle rear-view mirror phenomenon in an ordinary cycle, and we agree. But our own analysis of the current recovery suggests that the labor market will take a while more to follow historical metrics. We will cite, however, a recent paper by former Bank of England advisor and Dartmouth professor (among other things) David Blanchflower and Andrew Levin, currently at the IMF but moving to Dartmouth in the fall, in which they publish some new results on underemployment and wages drawn from state-level data. In a nutshell, they find that a decrease in broadly measured labor-market slack does not increase wage pressure if the level of slack remains high, and assert that starting the tightening process would be premature, and should be held off until labor market slack decreases significantly and inflation comes closer to the FOMC’s goal of 2 percent.

Considering its recent performance (see below), if the stock market can’t withstand up to a 1% increase in the policy rate, we’d better find that out. 

The environment of super-low rates and high liquidity has encouraged all kinds of financial machinations. The withdrawal of that indulgence, we hope, will encourage economic actors to focus  on more productive long-term activity. We suspect that a move from abnormally low rates to still very low rates won't derail the welcome increase in capital spending we have seen recently. 

We support concern about the effect the first move may have on the improving job market–it still has a way to go. A recent SF Fed piece suggests a few more years to true health.  That makes it a real shame that there's little possibility of a productive fiscal compromise: a focused business and low-inflation friendly infrastructure/retrofitting project that could offset the potential pressure on our workforce as rate increases turn the corner. Not to mention some relief from the ever increasing flat tire rate caused by old degraded roadways.

What gas dividend?

Comments & Context

Oil (and gasoline) prices have crept off their lows; they’re up about 25% this year, though they’re still well below year-ago prices. But many analysts have been scratching their heads wondering why the decline in energy prices hasn’t done more to goose up retail spending. It may be that's not really much of a conundrum. 

PCE

Here are several ways of making that point. First, there’s the graph above, showing the annual growth rate in nominal expenditures on energy and all other forms of consumption from the national income accounts. If energy prices could add to or subtract from other forms of consumption, you’d expect the two lines to look more like mirror images. In fact, they move more together than apart. The correlation coefficient between the two series is 0.43. Lag the non-energy spending a year, on the supposition that there might be a delayed reaction time, and the correlation coefficient actually falls to 0.28.

Moves in energy consumption are overwhelmingly dominated by price. The correlation coefficient between the annual change in the gasoline price and energy expenditures is 0.97. (That’s made even more remarkable by the fact that more than a third of energy spending is on services like electricity, which are not priced like gas.) While people may drive more or less, or raise or lower the thermostat, based on price movements, it appears they don’t change their behavior all that much. They just take the price—and it doesn’t seem to have much effect on other spending.

One gets similar results looking at the BLS’s Consumer Expenditure Survey (often abbreviated CEX, so as not to confuse it with Current Employment Statistics, the official name of the payroll survey). The CEX can be volatile, and dependent on the unreliable memories of the surveyed in describing their spending habits, but it’s still the best household-based measure of expenditures we have—and it allows for fairly detailed demographic analysis.

Ene-Nonene-CEX

As the first graph above shows, we don’t find the mirror image between energy and non-energy spending growth that you might expect in the aggregate. The correlation between energy and non-energy spending growth is 0.28—positive, not negative. And the story is similar when you look at poor (bottom quintile), middle-income (third quintile), and affluent (top quintile) households. The correlation weakens the higher you go up the income ladder (0.52 for the first, 0.42 for the third, and 0.14 for the fifth), but they remain positive for all five. 

Not surprisingly, there’s a tight relation between income growth and spending. A regression of the annual change in non-energy spending in the CEX on the annual change in household income yields an impressively high r2 of 0.53. So the “mystery” of why lower energy prices are not translating into higher retail spending may not be a mystery at all. The best explanation of tepid retail growth is tepid income growth.

Hourly earnings: a longer view, but still weak

Employment & Productivity

 It wouldn’t hurt to have some longer-term perspective on what’s been happening with average hourly earnings as the labor market tightens. In a couple of words: not much.

Graphed below is a history of the yearly growth in nominal average hourly earnings (AHE)—since 1965 for production/nonsupervisory workers and since 2007 for all workers. (The all-worker series is fairly short, making long-term comparisons impossible, but it does track the production worker series pretty well.) Note that both series are very close to their historical lows.

  AHE-yty

For the year ending in February 2015, AHE for production workers are up 1.6%—half the 3.1% 1983–2015 average. That figure was a little lower in 2012, but it’s quite close to three earlier troughs. In May 2014, less than a year ago, the annual gain was 2.4%. So we’re down 0.8 on wage gains—coincidentally, exactly what the decline in unemployment has been over the same period (6.3% to 5.5%). Wages aren’t usually thought of as decelerating as the labor market tightens, but that’s what’s happened.

The slowdown in all-worker AHE is less dramatic, but February’s 2.0% annual gain is below the 2.4% average for the entire series, and 2014’s 2.1%—and considerably below the 2008–2009 average of 2.9%, when the economy was falling apart. 

So, really, weak wage growth looks to be more of a macroeconomic problem now than wage pressures.

 

 

Advanced industries

Comments & Context, TLR in the News

February 11, 2015

 The Brookings Institution just published a paper outlining the importance of our advanced industries (AI), and a call to do more to support its growth. To be classified as advanced  industries must spend more than $450 per worker on R&D, putting them in the 80th percentile, and the share of workers whose jobs demand a high degree of STEM (Science, Technology, Engineering and Math) knowledge must exceed the national average of 21%. Overall, industries so classed constitute our tech sector at its broadest: most manufacturing operations including aerospace, automotive, navigational instruments, medical equipment and supplies; energy including electric power, mining and extraction; and services, including architecture and engineering, data processing, satellite communication and scientific research. 

Between 1980 and 2013 such industries grew 30% faster than the rest of the economy, or 5.4% annually, but employment levels held basically steady, largely because of productivity advances. Since the Great Recession both activity and employment have been on the rise, the sector adding one million direct jobs between 2010 and 2013, with employment and output rates 1.9 and 2.3 times faster than the nation as a whole. And AI employment has a high multiplier: those 1 million direct jobs likely generated another 2.2 million outside the industry itself, which compares to about 6 million total jobs created nationally over the same stretch. Advanced services were responsible for 65% of the total AI jobs, with computer design by itself up 250,000. Industries building transportation equipment have been adding jobs, finally, after decades of losses. A lot of attention has been focused on contributions to national payroll gains made by AI subsector gas and oil extraction since the recession. 

A boost? Certainly, but dwarfed by the mothership.

 Each worker in the sector generates about $210,000 in value added, in comparison to the $101,000 average overall, which helps offset the sharp increase in wages. In 2013 the average AI worker earned $90,000 annually, close to twice the overall average. Between 1975 and 2013 inflation-adjusted AI wages were up 63%, compared with 17% outside the sector. Total AI employment was about 12.3 million in 2013, about 9% of total overall employment, with another 27 million jobs riding AI's coattails. Together with the directly employed, that’s about one-fourth of US employment. Surprisingly, about half of the workers in the sector hold less than a bachelor’s degree, and Brookings ranks it as an accessible field. And one where more workers are needed.

Computer systems design owns the largest share of total AI employment, 13.8%, followed by architecture and engineering, 11.0% and management and technological consulting, 9.6%. Between 2010 and 2013 fastest rates of employment growth took place in information services, 10%; railroad rolling stock, 8.4%; automotive, 8.1%, and gas and oil extraction 7.6%.

Advanced industries tend to thrive in urban areas: San Jose is our most advanced hub, with 30% of the workforce laboring in the sector, followed by Seattle, 16%, Wichita, 15.5%, Detroit, 14.8%, and San Francisco, 14%. San Jose, Detroit and Seattle have the broadest array of advanced industry. (We’ll be presenting more data on this in coming months.)

 We are losing international share, and there are a host of serious problems we like to nag about, but even in its somewhat weakened state our AI sector a powerful thing. And if we’re worried about productivity, as we should be, it’s a great place to start.

 

Current Quit Rate Consistent with 7.7% unemployment

Employment & Productivity

March 6, 2014

Lately we’ve been seeing the argument made that the labor market is tighter than it looks. The argument goes like this: while the decline in the unemployment rate may have been boosted by labor force withdrawal, many of the dropouts will never work again, so it’s wrong to adjust the official jobless rate, either statistically or mentally, to compensate for depressed participation rates. So January’s 6.6% unemployment rate, 0.1 point above the Fed’s long-standing trigger, is getting close to “full employment,” and it would be prudent to start thinking about raising the fed funds rate sooner than most market participants expect. Is there anything to this?

We think not. For one, a 6.6% rate is about two-thirds of a standard deviation above the 1948–2007 average of 5.6%, which is not trivial. It’s even further above, in absolute terms, its 2002–2007 average of 5.3%, an expansion that was far from robust. It may be close to the Congressional Budget Office’s estimate of the “natural” rate of 6.0%, but as we showed last month, there’s nothing very scientific about these estimates; just six years ago, the CBO projected that the natural rate would be 4.8% right now.

A more subtle version of the argument looks at sectoral unemployment rates and finds some getting awfully close to full employment. We have a hard time seeing that. Graphed at right are unemployment rates by major sector compared to their 2000–2007 averages. In only one sector—manufacturing—is the January 2014 unemployment rate close to its average, though it’s 0.1 point above. Next closest is finance, 0.8 point above. The others are 1–2 percentage points above their average. A nice theory, but it just doesn't hold water.

Relatedly, some analysts are detecting wage pressures under a placid overall average—a 1.9% gain for the year ending in January. One argued that weakness in financial sector pay is dragging down the average. But if you do a weighted average of hourly wage growth excluding that sector, you still get 1.9%. Most major sectors, accounting for 72% of total private employment, exhibit wage growth below their 2007 average. In fact, wage growth is slower than it was a year ago—and that’s true of sectors accounting for 68% of private employment. It’s hard to see any tightness here either.

quit rates

Another place to look for signs of labor market tightness is in the quit rate. If workers perceive jobs as easy to get, they’re more likely to quit on their own. And, short of that, they’re more likely to demand raises from employers eager to keep them. But currently the quit rate is low by historical standards.

The BLS started publishing the quit rate in 2000. Since the quit rate tracks the number of those unemployed 5 weeks or less very closely, we used that series to estimate the quit rate going back to 1967. (Where the two series overlap, the fit is very tight—an r2 of 0.93.) December’s 1.7% rate is well below the full series’ 2.1% average. It’s also well below levels seen close to previous business cycle peaks, like 1979, 1989, 1999, and 2007.

The quit rate moves generally in line with the unemployment rate. You can “predict” the unemployment rate with decent accuracy with the quit rate, in fact. But as the graph on the bottom of this page shows, the unemployment rate associated with the December 2013 quit rate is a full point above its actual level. Or, putting it more bluntly, workers are acting as if December’s unemployment rate were 7.7%, not the 6.7% it actually was. If the job market were tighter than it looks, we’d expect a much higher quit rate.