Anyone looking for a job in 1940 couldn't email a headhunter or search the web for openings. They could, however, bring a mimeographed copy of their resume to the interview. If hired, odds were they became lifetime employees of the company, most likely a heavy industrial manufacturer.
Over six and a half decades, Americans have experienced a sea change in how we look for work, where we work and how often we find new work. We have progressed into a wired, upwardly mobile, flexible workforce. Small businesses, self-employment and independent contracting have become the hallmarks of our entrepreneurial, innovation-driven economy.
With such a drastic transformation, one might expect the way we measure employment has evolved too. Yet our most frequently cited survey of job creation remains mired in a Depression-era mind-set and research method. The Bureau of Labor Statistics' Current Employment Statistics survey, or the payroll survey, tracks payroll employment by surveying established businesses. This results in monthly job creation numbers.
One other important jobs survey is the Current Population Series, or the household survey. The household survey tracks all types of employment -- from someone who holds a lifelong job at a big business to someone who just became their own boss -- and it produces the unemployment rate. The household survey tends to be much more volatile than the payroll survey, which is a main reason it is used less frequently.
In fact, the public and private sectors have historically overlooked the household survey and relied on the payroll survey to gauge national job growth. When we look back to the pre-WWII economy, favoring the payroll survey makes sense. At the time, there was relatively little variety in the types of jobs available or the way they were created. Familiar companies, such as U.S. Steel and General Motors, dominated national employment.
Today, the employment landscape is entirely different. Just look at Southern California, with its biotechnology facilities, independent IT contractors, and small, specialized consulting firms. Yesterday's start-up is today's big business, and today's brainstorm is tomorrow's start-up. It's not surprising, then, that the payroll and household numbers portray quite different results.
The disparity between these jobs surveys became particularly apparent throughout the early stages of the post-recession recovery in 2002 and 2003. While the payroll survey lagged for months, the household survey demonstrated a strong and growing workforce, where self-employment accounted for a third of all job creation.
Following the end of the recession in November 2001, job creation in the household survey had rebounded by the following May. Although there were some ups and downs in the ensuing months, the household jobs numbers never again dipped below the November 2001 level. By November 2003, more than 2.2 million net new jobs had been created, and the pre-recession jobs number had been surpassed.
By contrast, the payroll survey did not demonstrate net job growth until August 2003, and didn't return to the November 2001 level until April 2004, nearly two years after the household survey had caught up. And the payroll survey's pre-recession jobs number was not surpassed until February 2005. This prolonged lag in the payroll survey's job creation numbers led to claims of a "jobless recovery." While every other major indicator of economic strength surged forward, from GDP to productivity, the payroll survey persisted as an anomaly of negative news. In an already dynamic economy, the increased churn created by economic expansion only highlighted the growing inadequacies of a Depression-era payroll survey. Using mid-20th Century methods to take a snapshot of the 21st Century employment picture simply does not work.
To launch an overhaul of our jobs surveys, Majority Leader John Boehner and I introduced H.Res. 14 last year. The resolution called on BLS to review and modernize the way we collect jobs data. In response, the BLS conducted a report that analyzed the two surveys and evaluated options for change. While the report stopped far short of proposing a complete reform of the surveys, it did acknowledge that a growing discrepancy exists between the two numbers and determined that further analysis is necessary. I am pleased that BLS has taken this important first step. But it is only that. We must continue to push for reform so that our jobs surveys effectively track job creation.
Policy makers rely on accurate economic data to draft effective legislation and businesses need the right numbers to plan for their future. In an economy where the only constant is change, unreliable numbers will result in off-target legislation and poor business decisions. A modern economy needs modern statistics.
