Friday, May 7, 2010

Mixing can lead to headaches. How much did employment increase in April?

My father always drank Vodka on the rocks. He warned me that mixing liquor with anything except ice cubes would produce undesirable results the following day. While I mostly drink bourbon or sour mash, I generally follow his advice. And as a good son, I try to follow it regularly.

I wish the press had learned from my Dad. On Friday the US employment data for April was announced with most headlines noting that payrolls rose by 290,000 while the unemployment rate rose from 9.7% to 9.9%. This seeming incongruity was swept away with a passing reference to growth in the labor force. None of this is incorrect but unless you are a macro-nerd and go to the appropriate place -- -- you could be seriously misinformed.

You will not learn from most of the headlines that employment actually increased by 550,000 persons in April. The issue and confusion – did employment rise by 290,000 or 550,000 – seems worth discussing further. One television commentator was trying to calculate how long it would take the unemployment rate to reach an acceptable (lower) number if job growth continued by 290,000 per month. Of course, she was using the wrong number – it would take about half as long as she calculated if employment continues to rise at 550,000 per month…

So what is all the fuss here? The source of the problem is that the Bureau of Labor Statistics of the US Department of Labor publishes two surveys each month – the Current Population Survey (which tells us employment and unemployment numbers) and the Current Employment Survey (which tells us the number of jobs created). Or more briefly, these are called the Household Survey and the Payroll Survey.

The Payroll Survey is a measure of jobs but not employment. The monthly sample covers 390,000 nonfarm establishments – government and private. These nonfarm establishments tell the government the number of jobs on their payrolls. If a person works for two firms – one full-time and one part-time – they would be counted twice. Anyone self-employed or working for a company categorized as agriculture, would not be counted in the Payroll Survey.

The household survey involves about 60,000 households. They are asked about the labor force status of each household member. The survey classifies each person in the household in one of three ways – (1) in the labor force with at least one job (employed), (2) in the labor force and looking for a job (unemployed), or not in the labor force (16 years old or older and not looking for a job). If you want to know exactly how they categorize people through the survey, go to

The household survey counts you as employed or not – so it underestimates the number of jobs. It also includes self employed persons, people who work in agriculture , and unpaid members of a family business. Because it is a sample and prone to errors, the BLS explains that it takes a change of at least 436,000 to be statistically significant. So the 550,000 has meaning for April.

The April Household Survey estimated that the labor force increased by 805,000. That means that 805,000 people wanted a job and 550,000 of those people found them. It also means we added 255,000 (=805,000-550,000) to the pool of unemployed. Thus the April unemployment rate (all people unemployed of 15.2 million divided by all people in the labor force, 154.7 million) rose to 9.9%. But surely the good news is that 550,000 persons found work and there were 255,000 encouraged enough by current conditions to move from not looking for work to looking. Compare that to most of 2009 when the civilian labor force was declining.


  1. Hi Larry,

    Your last paragraph confused me. Specifically you say "That means that 805,000 people wanted a job and 550,000 of those people found them." Seems to me that should be 805k MORE peeps wanted a job for a total of ~15 million peeps who wanted a job.

    I also wonder if these numbers need to be seasonaly adjusted. Lots of kids were planning for graduation and entered the job market; not because they were encouraged by prospects of employment but because they could not get in grad school. ;)

    All kidding aside you seem more optimistic than I am about the economy. Problems in Europe have already made the dollar stronger and the IMF may want more US tax dollars for Spain and who ever else. Greece was a small fry economically compared to who may be next. I spend way too much time watching the stock market and am not happy with what it is going on there.

    Bottom line is that long term govts cant continue spending more money than they take in from taxes and expect the economy to improve.

  2. Mike,

    You need to keep the stocks and flows straight. I did say in the post that the total number of unemployed in April was 15.2 million. But the changes during the month had 805,000 entering the labor force and of those, 550,000 found jobs. The numbers are seasonally adjusted. There is nothing in this post that exudes optimism. I just pointed out a confusion some people have in the monthly data. Reread my Whack-a-Mole post for my assessment of the economy. I don't think that is very optimistic.

  3. What I am confused about is that of the 550k peeps found jobs you seem to be saying they all came from the 805k that entered the labor force; and it seems to me they could have come from the 15 mill unemployed or the new 805k; or most likely from both the already unemployed and the new peeps entering the labor force.

    Maybe I just need another cup of coffee; or maybe some vodka and ice.

  4. This comment has been removed by the author.

  5. I made a critical typo some I am adding another comment to replace the one above.

    What I said about stocks and flows is the key to the answer. The 805,000 is the net addition to the labor force in April. Some people left the labor force and some entered. The net flow to the labor force was 805,000. The same net idea applies to both employment and unemployment. On net, employment increased by 550,000. On net the labor force increased by 805,000. When the dust settled there were 255,000 more people in the group of unemployed. Yes, some of the 550,000 employed increase came from new entrants to the labor force and some came from the unemployment pool.

  6. The components of how these numbers are calculated by the government don't seem to come from a reliable representative sample in the first place so what is the usefulness?

  7. Who says they are not reliable samples? There are two different numbers and they describe slightly different aspects of employment. We have many indicators of inflation -- it is okay to have more than one indicator for a macro concept.