Yesterday, the Wall Street Journal reported on the growing controversy of a chart produced by Josh Bivens and Lawrence Mishel of the Economic Policy Institute and its meaning. Let's dive right into the center of the storm.
In case you missed it, there’s a chart making the rounds that has come to represent, for some, all that is wrong with the American economy.
The top line shows worker productivity growing by 72.2%, or 1.33% per year, between 1973 and 2014. The bottom line shows median workers’ hourly compensation increasing by 9.2%, or 0.2% annually, over that same period. The gap between them more or less symbolizes the big empty space in workers’ wallets.
The chart, part of a broader research series from the left-leaning Economic Policy Institute, has struck a chord. Hillary Clinton tweeted a version of it following her first major economic policy speech in July, and has clearly leaned on EPI research in calling for policies that boost wages.
But the chart—and its data, methodology and conclusions—has become a flashpoint.
How one interprets the mass of historical data we have on workers’ productivity, wages and growth has profound consequences for which economic policies one might pursue. Most economists agree that productivity is essential for raising living standards, and that the superrich have seen their incomes skyrocket. The devil is in the details.
The economists behind the chart have now come out with guns blazing, releasing a vigorous defense of their work that decries critics’ attacks as “baseless.” The briefing paper by Josh Bivens and Lawrence Mishel, “Understanding the Historic Divergence Between Productivity and a Typical Worker’s Pay,” calls wage stagnation for the majority of American workers “a bald fact.”
So is it?
The WSJ article then goes on to take in the viewpoints of a number of other economists, including such players as Scott Winship, Robert Gordon, and Robert Lawrence, who point out a number of the problems with the data and analysis used to produce the chart.
But there's a more fundamental problem with both the data and the analysis, which these critics have only barely touched upon. Bivens and Mishel got the math wrong. Badly.
Not their actual calculations, mind you. Their problem lies deeper in a seriously flawed methodology - one that students in hard sciences, such as physics or engineering, are specifically trained to avoid.
But you don't have to take our word for it. Russian geophysicist Ivan Kitov was intrigued enough by Bivens and Mishel's results that he reviewed their work the same way any Doctor of Physics and Mathematics might review the work of a student submitting a term paper or a final exam in an undergraduate level class.
As you read the following, keep in mind that English is not Dr. Kitov's first language and that being Russian, he doesn't have any stake in the political arguments involved, which would only apply to the U.S. He may be as close as we can get to a genuinely neutral arbiter - although as you'll find out, not one who tolerates incompetent analysis well.
In terms of physics, democratic (might be not only) economists are ignorant. I would recommend learning some basic notations from thermodynamics before using any specific units. The physical concepts of ensemble and closed system seem to be too difficult for economic science.
My story is simple. Through "Economist's View" I found a graph in The Fiscal Times comparing the evolution of productivity and hourly compensation. This figure is not too complicated; it's rather too misleading.
Democrats claim that money leak from "workers" to "wealthy", whatever it means. This is not true; this just demonstrates the (hopefully not deliberate) misuse of simple notations. Roughly, hourly compensation is calculated as a ratio of total wage and salaries and total hours worked. Indeed, if to ignore the increase in employment/population ratio since 1950, and especially since the late 1960s, one gets something as shown in the figure below. Real GDP per capita and labor productivity ($ per hour as reported by Total Economy Database borrowed from the Conference Board website) follow similar paths. Wages and salaries (as published by the BEA) divided by employment (same database) deviate from these two curves since the 1970s, as in the above figure.
The next graph shows the ratio of employment and total population since 1950. This ratio has been increasing from the 1960s. Effectively, more and more people are involved in real economy, but unfortunately for them they share the same real GDP. On average, one person gets smaller and smaller share of the cake - and we see that compensation per hour increases slower than GDP per capita and productivity.
When compensated for the difference in the total population and employment growth, the green curve is back to its true position. Rich do not rob "workers". Instead the employment has been increasing over time. In that sense, declaring the decrease in hourly compensation as evil, democratic economists are strongly against increasing employment, i.e. against workers.
Summary
Never normalize values to fluctuating portion of a closed system. This always gives a biased (wrong) result. Also, it is always a formal mistake and negative mark on exam (in physics).
It is also a prime example of what Paul Romer has called economics' "mathiness" problem, where highly deficient analysis is too often generated for the sake of pursuing a highly politicized agenda in the media, where journalists who lack sufficient training in the sciences are exploited into advancing the biased analyst's political agenda.
In this case, Bivens and Mishel's fundamental error has not just created a controversy within the field of economics, it is actually setting back positive progress in the field. Worse, the policy prescriptions being advanced to "correct" Bivens and Mishel's identified "problem" by U.S. policy makers buying into their fundamentally flawed analysis would, if adopted, actually set back the economic interests of the very people the political "solution" claims to help.
All for a story that doesn't stand up to scrutiny. We can do with a lot less of those in this political season.