Today Bureau of Labor Statistics (BLS) announced unemployment numbers for September. Drop in unemployment rate stirred discussions as it makes Obama’s hand stronger in the presidential race. Some went as far as calling the numbers massaged but majority of the skeptics claimed that the decline was due to increase in part-time jobs and as part-time jobs get cut first when economic conditions worsen “It’s an employment recovery built on thin ice”.
Let’s look at some of the claims and the data
According to Bloomberg Business Week “Some 582,000 more Americans, the most since February 2009, were working part-time last month because of slack business conditions or because those jobs were the only work they could find, according to Labor Department estimates.”
Labor Department estimates tell a different story. According to the data, 657,000 fewer Americans were working part-time for what BLS calls “economic reasons” in September 2012 compared to September 2011 and it was certainly not the most since February 2009.
According to Boston Globe, upon release of BLS data, John E. Silvia, chief economist for Wells Fargo & Co. said “September’s rise in part-time employment was due to college students returning to school and taking part-time jobs.” and it was the gain in part-time work that pushed U.S. jobless rate down.
According to the data, part-time jobs (secondary axis) are in decline in the past few months following a sharp increase between February and April. There’s a clear increase in full-time jobs (primary axis) in the recent months. Looking at this graph, it is hard to attribute the unemployment rate drop between August and September to part-time jobs.
So at least based on BLS household survey data, unemployment rate drop in the last two months is clearly because of the increasing number of full-time jobs. However to understand unemployment, it is important to look at how the rates are calculated.
The Labor Department uses two surveys, one to estimate payroll employment, and the other of households, to estimate unemployment. They both have their shortcomings and don’t always correlate with each other.
Household Survey, involving just 60,000 households, is subject to a relatively large sampling error. Also in this survey, if a person has worked just one hour, they are counted as employed. Even people who did unpaid work in a family-owned enterprise (as long as it is more than 15 hours) are classified as employed. The payroll survey sample does not include new firms immediately. They are incorporated with a lag.
Another issue is the definition of labor force. Labor force measures are based on the civilian non-institutional population 16 years old and over and is made up of the employed and the unemployed. Persons are classified as “not in the labor force” if they do not have a job, and have not actively looked for work in the prior 4 weeks of the survey. So it is very easy to be unemployed but not counted as such.
Since U3 is how “official” unemployment rate has been calculated for years, people often refer to the changes of this parameter over time when they talk about unemployment. It makes sense for an apples-to-apples comparison, but there are other approaches considered to be closer to real unemployment such as U6. U6 considers involuntary part-time workers (those who took part-time jobs because they couldn’t find full time jobs) and marginally attached/discouraged workers (unemployed people who haven’t searched for jobs in more than 4 weeks) as part of unemployed labor force.
With either approach, unemployment appears to be in decline while with U6 we observe a considerably higher unemployment rate.
Looking at interest in the terms such as food stamps and part time jobs on Google, it is hard to talk about a recovery yet. Maybe U3 needs some criticism after all.