News reporters--unless you’re prepared to read and understand the entire news release, including the “Technical Notes,” don’t write about productivity and wage statistics reported by the Bureau of Labor Statistics.
Even James Surowiecki in the March 2, 2009 issue of the New Yorker wrote a piece that appears to be based in part on the February BLS Productivity and Costs news release. He is generally an incisive and well-informed business writer.
He had seven points to make:
• While millions of jobs have been lost, average hourly wages have increased.
• Employers don’t cut wages because wage cuts “wreck employee morale and eat away at productivity.”
• Productivity has risen at “an impressive 3.1 per cent.”
• Workers get paid more the more productive they are.
• Companies used to “hoard labor” in downturns, but with the just-in-time economy, they are quicker to fire people.
• Companies are obsessed with doing more with less.
• We need the stimulus package because prospects are bleaker for those out of work.
That sounds eminently logical and based on data. But wait a minute—what data?
As you read the technical notes of the news release that start on page 8, you see words like adjusted, estimated, sampled, imputed, based on ratios, deflators, aggregated, NAICS, chain-type current-weighted index, assuming…
Your eyes have glazed over already, haven’t they? These are highly-derived, 20,000-foot level statistics developed in the 1930s and 1940s to help make economic policy decisions. (The effectiveness of such decisions may be open to question.) Each metric that feeds the aggregated data has its own derivations, uses, and margins of error.
The people who gather and analyze this data are not misleading anyone. They issue all sorts of publications explaining what they’re based on and how to read them. They continuously improve how they measure things, and improve accuracy. They publish retrospective critiques of their own forecasts. They’ll talk to you if you call them with a basic question and they’ll answer your e-mail.
Surowiecki started with “average hourly wages,” which have indeed gone up at the same time the number of jobs has gone down. He suggests that, with productivity rising, the best and most productive workers are keeping their jobs and the slackers are being let go. But what are “wages” and “hours”?
“Wages” means total compensation paid to all people working in the business sector of the economy, which excludes a few categories like government workers and employees of nonprofit organizations. That total compensation includes the millions paid to some people, and the pittances paid to others. When lower level employees are let go, but the big shots stay, the average is going to be distorted. Total compensation includes the cost of benefits, including health insurance and workers comp. Speculating on average pay for the average worker based on this data is not a good idea.
“Hours” is a calculation of all hours worked by all people in the business sector workforce. It excludes vacations and other time off, based on surveys of company policies. It can include overtime, or not. (As the workforce shrinks, how many people are working unrecorded overtime to keep up?) It’s an approximation. You can compare this year’s figure to the same figure five years ago and draw some conclusions about the economy as a whole. You can’t draw conclusions about how much time an average worker is putting in.
“Average hourly wage” is an approximation divided by a guess.
Do employers avoid cutting wages? Yes and no. If there are any new hires, are starting salaries lower than before? Are wage increases smaller? Are unions being asked to make concessions? Are benefits being cut? Are employees contributing more of their own money to get benefits? What about bonuses?
The biggie—productivity. Has it risen an “impressive 3%”? Do more people have their noses to the grindstone? Are American workers the most hard working and productive in the world? We really like to think so—and most people do work hard. But that has nothing to do with “productivity” as reported by the BLS.
“Productivity” is just another calculation. Gross domestic product—all the stuff sold for consumption in a given time period—divided by the number of hours worked. (Note that it doesn’t include all the stuff produced that got left in a warehouse somewhere.) In the last reporting period, the economy made and consumed slightly more stuff, while fewer hours were worked—since fewer people were working—so “productivity” is up. But “productivity” is just an estimate divided by an assumption.
You can speculate on a lot of situations that could feed productivity—for example, an unemployed person works no hours but still consumes what somebody else makes. But long term productivity growth, what we like to point to when we want to show how hard Americans work, is largely driven by the work done by the information systems and machinery that used to take many hours of labor to do. That’s the dirty little secret. You can’t say that anyone is working harder at any time compared to any other.
In fact, right on page 9 of the news release, if you bother to read that far, the BLS says:
Although these measures relate output to hours at work of all persons engaged in a sector, they do not measure the specific contribution of labor, capital, or any other factor of production. Rather, they reflect the joint effects of many influences, including changes in technology; capital investment; level of output; utilization of capacity, energy, and materials; the organization of production; managerial skill; and the characteristics and effort of the work force.”
You can’t speculate on less productive people being fired while more productive people keep their jobs, as Surowiecki does.
Next question: Does a “just-in-time” economy cause companies to fire people more quickly? I have no idea what this is supposed to mean.
Are companies obsessed with doing more with less? Yes, but they don’t usually go about it in any effective way. You can randomly fire thousands of people and hope that those left will magically be able to think of ways to get all the work done. Or, less likely, you could be going about continuous improvement in a systematic way and letting attrition run its course if you need fewer people, or hiring on an agreed-upon contingent basis and do without them when things are slow.
Do we need the stimulus package to include extended unemployment benefits? I think that’s true, but that doesn’t have anything to do with the BLS cost and productivity report.
I’m sorry to pick on Mr. Surowiecki. He hasn’t done any worse than 99% of the media.
However, we have a responsibility too. As consumers of economic news, we have the obligation to dig down to the roots of assertions from time to time. With the web, all we need is to take a little time and read the first nine pages of a press release. Just once.