The most closely watched and widely cited economic statistic in the world may be the unemployment rate. It first came into use in the 1930s, and its definition remained basically unchanged since. However, there is a growing realization that it has become an unreliable economic indicator, and it is time to find a better way to measure the health of the U.S. labor market.

Policymakers are starting to take note, with Rep. Michael Fitzpatrick, R-Pa., introducing legislation that would change how the Bureau of Labor Statistics measures unemployment. Even one of the last holdouts, the Federal Reserve, now seems to be moving toward dropping it as a quantitative target in favor of using a broader range of additional economic data.

The intent of the unemployment rate is to measure slack in the labor market, calculating both the demand for labor (i.e., employment) and the size of the available labor force (those who either have a job or want one). Right now, both measures are flawed, leaving the Fed, Congress and others essentially flying blind as they make critical economic policy decisions. The threshold for "employment" in any particular month is quite low: An individual needs to complete just one hour of any type of work for pay or profit.

Even without earning a dime, people who are temporarily absent from work or spend at least 15 hours working for no pay in a family business will still count as employed. This very crude measure is no longer consistent with the actual experience of those who lose a job. A significant amount of research now shows that most long-tenured workers who become unemployed will, even after re-employment, remain underemployed and have lower earnings for as many as 15 years.

Part-Time Nation

However, the BLS still counts experienced engineers doing nothing more than a couple of hours of paid yard work for a neighbor as employed — just as if they were continuing to work in their original occupation at the same pay. The current measure of employment does not capture underemployment with respect to occupation, salary or hours worked. For example, we know that there are currently about 8 million people working part-time only because they can't find full-time work.

More important, there is considerable anecdotal evidence that millions count as working — but are not in their normal occupations and not earning their normal wage or salary. A new study by the Federal Reserve Bank of New York finds that recent college graduates are settling for more low-wage and part-time positions previously held by those with a much lower educational attainment.

Measuring the available supply of labor is equally problematic. The current measure of the labor force includes those who are employed, in addition to unemployed individuals who are currently available and actively looking for work. To maintain unemployed status, individuals must contact employers directly, have an interview, ask friends or professional organizations about job opportunities, send out a resume or fill out an application each month.

Given the historically high number of long-term joblessness, it's no surprise that the share of out-of-work people who aren't considered active enough in their search has surged since the Great Recession. This — not rising employment — is the main cause of a drop in the unemployment rate since the end of the recession.

Declining Workforce

While 10.5 million people now count as unemployed, 91 million jobless individuals are not considered part of the labor force. Many of these individuals want work, and would likely be working if the economy were stronger. A better measure of the labor force would give a much clearer picture of economic health.

For example, the share of the population in the labor force should not decline during a recession. Every month many more people move into and out of employment from outside of the labor force than from unemployment, so the unemployment rate needs to give a better indication of who is close to returning to employment or will return once the labor market improves. For the BLS to produce a more comprehensive and accurate analysis of the labor market, new research should develop a better measurement of unemployment that accounts for underemployment as well.

Although the BLS currently collects data on those working part-time and those too discouraged to look for work, this isn't good enough. It should also factor in data related to whether individuals are employed within their usual occupations and earning their usual salaries. 

BLS should also study how many job searchers it excludes from the current unemployment rates simply because they use newspaper and Internet job searches to look for work. More broadly, BLS should think about moving from the "current, active job search" criteria for unemployment to a more modern definition that better identifies the permanent labor force.