Time to say goodbye to the unemployment rate

This week, as it does every month, the Bureau of Labor Statistics published its monthly employment report. As I wrote in The Wall Street Journal, we pay inordinate attention to this national number at considerable collective expense. While the statistic still has some value, it is time to stop focusing so relentlessly on the number and grapple instead with the complicated tapestry of employment today.

It isn’t just that the number is a statistical creation that involves both substantial estimation and frequent adjustments. It is that, and indeed, it is one national indicator that is subject to particularly dramatic revisions – none of which get the attention that the initial release does. If it’s reported in one month that businesses added 150,000 jobs to payrolls, that becomes a talking point. When the same number is revised down to 75,000 or up to 200,000 a month later (as often happens), it is a footnote. There’s yet to be a headline saying, “Jobs Picture Quite Different Than What We Said Last Month Because of Revisions.”

The unemployment rate is a recent invention designed for limited purposes, yet it has come to assume totemic status in a way that makes it almost impossible to have a cogent discussion of labor in the United States (or anywhere else in the world that keeps similar numbers) and design meaningful public and private responses to our challenges.

During the 2012 presidential campaign, it was widely trumpeted that no president had ever won re-election with an unemployment rate higher than 7.2%. That was said so often and by so many it was taken as a truism. Yet Barack Obama was returned to the Oval Office with the rate at 7.9%. Had some deep historical pattern been broken? Was his victory some deep anomaly? No.

Technically, the statement is true. But it is based on the flimsiest of foundations. There was no official unemployment rate until the late 1940s. The U.S. government only seriously began to define unemployment and measure it because of the Great Depression. As the crisis hit suddenly in 1929, the Hoover administration had no way to know just how bad things were. The result of these early efforts was to show that the disruption of the Depression was substantially worse than suspected, which then gave Franklin Roosevelt a powerful tool for victory in 1932. Hoover, a devotee of scientific management of government, was hoist on his own statistical petard.

It took more than a decade for methods to be refined, for sample sets to be introduced and survey methods improved. It also took years to decide what, precisely, was meant by unemployment, leading to the definition that still stands – namely, you are looking for work and cant find it, rather than simply not having a job that pays a wage. Not until the 1950s, in fact, did the BLS make any fanfare of its monthly numbers, and not until well after that did any media or politicians pay much attention to what these reports said.

So in framing the election of 2012 as hinging on the unemployment, almost everyone appears to have forgotten the following: at most, there have been sixteen elections when there even was an unemployment rate, from 1948 through 2012. In that period, only five presidents were re-elected, while two (Jimmy Carter and George H.W. Bush) lost their bid for a second term.

The statement that “no president has been re-elected with an unemployment rate of 7.2%,” therefore, should have been, “In the seven instances that we know of, no president has been re-elected with an unemployment rate of 7.2%.” To say the least, that would have lacked a certain punch.

But it would have the virtue of being true and thereby exposed the grand illusion of the unemployment rate and indeed of all of our leading statistics and indicators. We treat them as absolute markers, and reliable guides to our world. They are not. They are numbers that we invented remarkably recently and that then assumed a place of importance in shaping our sense of the world out of proportion to what they can and do measure.

A thousand years from now, if we are still keeping these numbers (doubtful), we will have more than enough data and patterns to make definitive statements about probabilities, correlations and possible causations. With less than fifty years of data, we are nowhere near that. The same could be said of establishing some relationship between jobs say and government spending. The great fallacy of the American Recovery Act of 2009 (better known as the stimulus bill) was that you could calculate precisely how much spending you’d need to produce 3.5 million jobs. The belief that $787 billion was the right number came from the formulas based on the limited amount of data over a few decades. Maybe there will come a time where we can calculate the right amount of spending. For now, we just don’t have enough information to make such conclusions, yet we do and do so in ever more aspects of our lives, from government to businesses to individuals.

Not only is there insufficient data to make sweeping conclusions, but there is also a major issue with the idea that there is a national unemployment rate at all. There isn’t. There is no one unemployment rate that encapsulates the national picture. If you are a college-educated woman, you unemployment rate is less than 4%. If you are an African-American male with a high school degree or less, the rate is well into the mid-20s. If you live in Oklahoma or North Dakota or other places in the shale oil belt, there has been no unemployment crisis in the past five years and that rate has consistently stayed below 5%. If you live in the areas hit hardest by the bursting of the housing bubble – Central California, Greater Phoenix, Las Vegas, vast swaths of Florida – or the areas decimated by the shifting sands of the auto industry, the unemployment rate has frequently been above 10%

There is, therefore, no unemployment rate, at least not as currently understood. There are multiple rates, by race, by gender, by geography and above all by educational attainment. When people talk of an unemployment crisis, it would be more accurate to speak of an education crisis or a crisis of men whose skills are mismatched to the jobs of today. It would be more accurate to speak of a jobs crisis in specific regions of the country, or for specific industries such as heavy industry. Yet, we maintain the collective fiction that one simple average accurately captures a multiplicity of realities.

By pretending that there is a meaningful rate, we then adopt policies that are bound to fail. Take the Federal Reserve. In the 1970s, Congress amended the mandate of the Fed to include achieving “maximum employment.” But if there is no unemployment problem per se in Tulsa or among college-educated women, then keeping interest rates low does them no particular good. On the flip side, if African-American males without a college degree are at a severe skills disadvantage along with any lingering racial barriers, then it’s hard to see how a zero interest rate policy is going to move that particular needle.

Bill Gates has been banging his virtual drum that in truth, there is no unemployment crisis; there is an education gap and a skills gap that then manifests itself as an employment problem. There are also those regional divisions, and there is yet another challenge posed by the informational technologies that Gates helped unleashed, which make factory labor and assorted other jobs irrelevant. Hence why you can have a manufacturing output revival without a manufacturing jobs revival. Robotics and just-in-time factory floors render much of the manufacturing labor force unnecessary.

Policies to address the problems of employment would have to be freed of the national fiction of an unemployment rate. The result would be targeted approaches. The politics of that are thorny. Congress is especially bad at allocating resources selectively, except in the case of natural disasters. Crafting legislation that privileged Detroit, Phoenix, central Florida, upstate New York, and so would find hard going.

But such targeted crafted legislation would be far more efficient and cost effective. Spending tens of billions of dollars a year on unemployment benefits is perhaps the least effective way to deal with the problems. It has a “let’s throw money at the problem and hope it goes away” tone to it, and relies entirely on the belief that once this thing called the economy gets humming again, employment will just materialize. Given the changes in labor and skills, that is, at best, unlikely.

The unemployment rate fixation represents an obstacle. It stands in the way of creative and innovative approaches, ones that involve government and ones that don’t. It locks us into a collective fiction that employment is a national problem with national solutions. The first step toward more dynamic approaches would be to wean ourselves from the unemployment rate. Time to say goodbye, and to plunge into the wealth of information about the world we actually live in and not the one that this number, and so many others, say that we do.

Zachary Karabel is the author of The Leading Indicators: A Short History of the Numbers That Rule Our World (just published by Simon & Schuster), from which this essay is adapted). He is also Head of Global Strategy at Envestnet. A shorter, similar version of this article first appeared in the Wall Street Journal (February 8, 2014).

Graphic: Jakee Zaccor / LinkedIn Pulse