AS THE ECONOMY TURNS

READING THE JOBS REPORT

Dr. William Shingleton

June 9, 2006

Last time we met we offered some guidelines on what to watch for in the maze of economic reports flowing out of Washington. This week, we thought we should follow that up with a more detailed look at what is probably the most watched report of all, the EMPLOYMENT SITUATION SUMMARY http://www.bls.gov/news.release/empsit.nr0.htm that came out last week. [NOTE: The BUREAU OF LABOR STATISTICS (BLS) has changed the name from the “Employment Situation Report” to the “Employment Situation Summary” because it is applying to the FDA to register the report as a non-addictive sleep remedy. In trial studies the number of people who can actually read the report from beginning to end without falling asleep is five.] The headline news for Wall Street was that the number of non-farm jobs in the economy had only increased by about 75,000 from the previous month. In their typical bi-partisan fashion, Republicans concentrated their attention on the word “increased” and Democrats concentrated on the word “only. Let’s take a look.

The major problem with the Employment Situation Survey is that it is actually an amalgam of two quite different statistical surveys, the HOUSEHOLD SURVEY, which is a telephone survey of 50,000 households in 792 sample areas and the ESTABLISHMENT SURVEY, which uses mail-in data from a sample of over 390,000 businesses employing over 47 million non-farm wage and salary workers, full or part time, who receive pay during the payroll period which includes the 12th of the month [SOURCE: http://www.bls.gov/lau/lauhvse.htm#hvse]. Because the data come from two different sources they seldom provide a perfect match and that allows whoever is using the data to make just about any political point he/she wants without actually falsifying the data. For the casual reader, the advantage of always having a focus on just one number is that he/she has something to steer his/her opinions, something like an economic North Star.

The Household Survey is actually a little easier to understand but the data is flawed. The BLS has the country divided into 792 sample areas, which works out to about 15 per state if we count the District of Columbia. They then call about 63 households in each sample area for information, which leads to the data flaws. Any household that has an unlisted number never gets called and is underrepresented in the survey. Since our guess would be that higher income households are more likely to bother with unlisted numbers this probably means that higher income families are underrepresented in the survey. Second, a number of younger households have given up on hard-wired phones altogether and gone to cell-phones, so that group is also uncounted in the survey. Another group that may be underrepresented in the survey is people who are out of work and can’t pay their bills. Our guess would be that some of them have had their phone service disconnected, making them very difficult to call. And then there is the most remarkable feature. This is supposed to be an EMPLOYMENT survey, isn’t it? Our guess would be that most of the employed people are not home to answer the phone when the BLS calls because they are working. Just a guess. [NOTE: This would be especially true because the BLS is a government agency and in a government agency when five o’clock comes, everyone goes home, unless they are a member of Congress. Your representative only has to work three days per week.] So who is home for the initial call? People who are out of work but not disconnected would be home to answer the phone, as would retired people and others not in the labor force. The BLS (hopefully) knows all of this and there are ways to RESTRUCTURE a sample (basically weighting the reporting demographic groups according to census data) but if they are only calling 63 households per sample area they don’t have a great deal of room to wiggle.

The other problem they have is the RELIABILITY of the data. Just think of getting an unexpected phone call from a complete stranger, “Hi, I’m from the BLS, not the CIA, and I need to know some important details about your economic life.” How do you know that it’s not someone trying to decide whether or not it would be worthwhile to break into your home? In today’s world you are foolish to share any economic (or personal) information with someone who calls you on the phone. To illustrate how people react to a random phone call, we sometimes have students from our statistics classes in our MBA program try the same kind of survey. The students are often stunned by the number of people who hang up (or worse), even after they identify themselves as students from Indiana-Wesleyan University. The information they get is sometimes so obviously unreliable as to be useless, which is one of the points of the exercise. It’s no surprise that ALAN GREENSPAN used to rely upon the alternative, Establishment Survey, instead.

The Establishment Survey has its own problems. The most obvious one is that it looks at NON-FARM EMPLOYMENT, rather than total employment, largely because farm-employment is so difficult to measure because of the remaining family farms and the temporary (and often illegal) workers hired for short periods. However, with farm-employment shrinking each year, this is becoming less and less of an issue. The second issue is more difficult to get a handle on, the survey asks questions of firms that have been around a while. However, new firms don’t even get into the survey until they become “established” and small businesses, where most of our employment growth is coming from, are therefore underrepresented. However, the survey has two very positive features. It includes 390,000 different businesses, rather than the 50,000 households of the Household Survey. In statistics, we usually find that larger samples are more reliable than smaller ones, although the difference is not nearly as large as you might believe. The other positive feature is that we are dealing with a business to government mail-in survey rather than a telephone survey. We would have to believe that the reliability of the data is much improved.

Now, let’s consider some of the more important numbers from the report. Our favorite, NON-FARM EMPLOYMENT, was up 75,000 from May. Since the number is from the Establishment Survey, it is an attempt to count the number of (non-farm) jobs, not the number of people who were working in May. Even if we ignore the statistical problems noted above, the numbers still would not match because some people hold more than one job. A number of Republicans noted, quite correctly, that May 2006 was the 33^rd consecutive month in which [non-farm] employment rose. [NOTE: For confirmation, see < http://data.bls.gov/PDQ/servlet/SurveyOutputServlet >] However, the Democrats were also correct when they noted that 75,000 was a fairly small increase for the month. [NOTE: Wall Street was expecting about 175,000 and we were expecting about 185,000. The shortfall in the number of jobs added to the gloom on the street. However if the survey had shown more jobs, that also would have added to the gloom on the street. The cheery optimists in the financial markets seem to have taken an early summer vacation.]

The most watched number for the general public seems to be the UNEMPLOYMENT RATE. The rate is calculated from the Household Survey, estimating the total number of people who are out of work but still looking for work, and then dividing that number by the total number of people in the labor force. The numbers worked out to an unemployment rate of 4.6 percent, which means that if you had a good representative sample of 1000 people who were members of the labor force 954 of them would have jobs and 46 of them would be looking for work. [NOTE: Why we have never been interested in the EMPLOYMENT RATE, currently 95.4, is an interesting question. Anyone know the answer?] Anyway, the 4.6 unemployment rate is a little below the average for the last ten years. [NOTE: The ten year high was 6.3 of June 2003 and the ten year low was 3.8 in April 2000.]

So with the unemployment rate a little low (and going down from last May’s 5.1) how could anyone complain? The problem is hidden in one of the lesser-lights of the Household Survey, the size of the labor force. If the size of the labor force was frozen in concrete then a 75,000 increase in the number of jobs, and the consequent drop in the unemployment rate, would have been great. However, in spite of the fact that this point that is underemphasized in almost every economics text we have ever used, the American economy is a dynamic growing system if we consider almost any reasonable span of time. For the labor force, what that means is that the total size of the labor force has been growing at an annual rate of 1.26 percent over the last 20 years. [SOURCE: Raw data: http://data.bls.gov/PDQ/servlet/SurveyOutputServlet] If the labor force had continued to expand at this rate during the Bush administration we would now need 141.4 million jobs instead of the 135.8 million we have, a shortage of 4.6 million. However, during the last five years our labor force growth has slowed down dramatically, to 1.05 percent per year, so we only need 139.9 million jobs. However, we only have 135.1 million jobs, so whichever way you want to count it, we are a little short in job opportunities.

The unemployment rare reflects two inter-related factors, the desire of employers to have more workers to produce their goods and services, which we call the DEMAND FOR LABOR, and the number of people in our society who are either working or who believe they can find a job if they look hard enough. The number of people in the worker and would be worker category is called the SUPPLY OF LABOR, once we allow that the number is influenced by what people expect to be paid once they find something. [NOTE: This is why we prefer the job count, rather than the unemployment rate as a useful measure. There are too many factors influencing the unemployment rate.] While the demand for labor grows as long as the economy expands faster than PRODUCTIVITY (output per worker) increases, the supply of labor is heavily influenced by EXPECTATIONS, what people who are not yet in the labor market think they are going to find if they decide to look for a job. As we have noted in previous reports, WAGES are stuck in neutral, barely rising over the last few years. In addition, the news is overloaded with stories about big companies laying off hundreds, or even thousands, of people. The small firms that add ten or twenty people at a time don’t make headlines, but the numbers indicate quite clearly that more people are being hired than are being laid off.

If people don’t expect to find a job that will pay a decent wage they won’t bother to look. In fact, economics says that it would be quite illogical for them to do so. As a result, our LABOR FORCE PARTICIPATION RATE, the percentage of the eligible population either employed or actively looking for work, which had rocketed up by 2.6 percent during the Reagan years, and then another full percent during the Clinton presidency, to 67.2 percent, has now dropped back to 66.1 percent, more than a one percent decline in a little more than five years. Our belief is that both Reagan and Clinton were effective optimists on the economy. Mr. Bush may be an optimist as well, but he has not been able to deliver an effective case on the economy to the American public. Again, it’s more of a guess than anything else, but both Mr. Reagan and Mr. Clinton had strong, effective Treasury Secretaries, who carried a good deal of the economic word to the public. Mr. Bush has had two abysmal choices for Treasury Secretary, which is why he is working on his third as we write. At least the new nominee, like President Reagan’s Mr. Regan and President Clinton’s Mr. Rubin, is from Wall Street. But that’s a story for another day.

Finally, once the job market gets rolling again, with rising wages and 220,000 new jobs each month, the unemployment rate may actually go back above 5.0 percent. That’s because, if people begin to believe that there are good jobs available, some of the people who have been DISCOURAGED WORKERS, to pessimistic to be considered actively looking for work, will return to the labor force, increasing the supply. Once they return and start looking they will be counted as unemployed, and the rate may go up. That’s one of the reasons why the unemployment rate is considered to be a LAGGING ECONOMIC INDICATOR, sort of like an economic writer you may know who does not get his material out on time lately.

Final notes: If you would like to be removed from the distribution list just send a reply on email. Back issues are available on my website <http://www.oocities.org/wsirius30/2cents.html>. The opinions expressed in these newsletters are those of the author. Comments, including suggestions for future newsletters, are always welcome.