If history were a walking path, we’d be standing at the foothills of a mountain. Looking back, we’d see the slowly changing landscape of human history represented by the valleys and hills we’ve traversed. Looking forward, we’d see a rapidly changing landscape of the future represented by the steep sides of the mountain we’re about to climb. The forces of history shaped the land we stand on: the valley and foothills behind by the agricultural and industrial revolutions; the mountain ahead by the coming digital revolution. In this metaphor the weather reports that we listen to each day are reports about the economy. The forces that shape the weather are not the same as the forces that shaped the earth, and the weather at the top of the mountain will be very different from the weather we experienced in the valleys below.
We recently weathered a huge storm called the great recession that, according to the National Bureau of Economic Research, ended over four years ago in June 2009. A mix of complex factors caused this storm, including government policy choices, high-risk lending and borrowing, and international trade imbalances. These observations are true as far as they go, but storms manifest differently depending on the terrain (valley, foothills, or mountain top), and the terrain is about to change very quickly. To make matters worse, we’ve failed to clean up the storm damage, a situation that economist Paul Krugman calls the mutilated economy. In a recent column, Krugman argues that current economic measurements “translate into millions of human tragedies — homes lost, careers destroyed, young people who can’t get their lives started.” He quotes a recent paper that argues that we have reduced our economic potential, and damaged our future economy, by tolerating high unemployment for so long. So, we’ve experienced a storm, we’ve damaged our future, and we’ll soon be hanging off the side of a mountain from an ice axe when the next storm comes.
If you’re like me, it is challenging to sort through all the economic reports. Economics is a complex topic and contradictory opinions about root causes and solutions abound. How do we sort through it all? I decided to download data from the U.S. Bureau of Labor Statistics (BLS) and do the analysis myself. The BLS issues a report on the employment situation every month, and the current edition can always be found here. I’m sharing my analysis because I want to show that any citizen can use and interpret the raw data themselves. If, like me, you’re not an economist, then I’m hopeful that this post will offer a reference for you to read and interpret the many articles about jobs and unemployment. Also, I’m providing a baseline for the future as I plan to post updates as the economic situation unfolds.
This is a long post with a lot of detail, so let me summarize the gist. The media has reported fluctuations in unemployment and job growth since the recession ended, so it feels like we are on a roller coaster. The recent jobs report was described in the New York Times as unexpectedly strong, with the pace of hiring greater than expected, yet the unemployment rate rose from 7.2 percent in September to 7.3 percent in October. At the same time, the size of the labor force dropped by 720,000, which brought the labor participation rate to a 35 year low of 62.8%. The news pushed the Dow Jones industrial average to a new high, yet at this pace it will take seven more years of monthly job growth to reach the pre-recession unemployment and participation rates. My claim is that the macro trend is clear and consistent: the percentage of the population that is at risk (AR), due to unemployment, involuntary part-time employment, or underemployment, has increased since 2007 and this trend will continue for the foreseeable future. In brief, more than one-third of the civilian non-institutional population is at risk by this definition. The pain won’t stop until policy makers, scientists, and economists figure out the nature of the new economy that is emerging in this century.
Bureau of Labor Statistics
Let me begin by walking through data from the U.S. Bureau of Labor Statistics. Here is a summary table of the economic situation as of August 2013. I realize that the data is slightly out of date, but it serves my purpose, and I’ll give an update in early 2014.
I’ll walk you through this row by row. The Bureau of Labor Statistics defines the population of interest as everyone over 16 who is not in prison or other institution. As shown in row #1, there are about 246 million such people as of August 2013. This population is split into two parts: everyone is either in the labor force (row #2) or not (row #9). Let’s first consider people in the labor force. This group is split into two parts: people are either employed (row #3) or not (row #8). The BLS calls the most common definition of unemployment “U-3”, which is the number of unemployed divided by the civilian labor force. In the news, this is simply reported as the “unemployment rate.” I give the numerator and denominator for this U-3 calculation in the rows labeled “U-3”, and show the rate as 7.3%. This unemployment rate tends to fluctuate up and down, and we need to carefully interpret it. Consider the people who are not counted. First, there are people who are working part-time, but want to work full-time. There were 7.9 million such people as shown on row #6. Second, there are people who are not in the labor force, but still want to work. There are 6.3 million such people as shown on row #11.
The BLS provides alternate measures of unemployment, and they call the broadest definition of unemployment “U-6”. This definition does not count all the people on row #11, but only those considered “marginally attached.” The BLS provides this very long, complex definition: “people who want a job, have searched for work during the prior 12 months, and were available to take a job during the reference week, but had not looked for work in the past 4 weeks.” I give the numerator and denominator for this U-6 calculation in rows labeled “U-6”. The numerator is the sum of rows #6, #8 and #13, and the rate is 13.7%. Note that the denominator is the sum of rows #2 and #13: those marginally attached are added to the total civilian labor force for this calculation.
Total At Risk Labor
These measurements are well-defined, but they don’t fully capture the way average people intuitively feel about the economic situation. I will use this same government data to create a different measurement that may do a better job. I’m calling this the “at risk labor force” — people who are either unemployed (want a job), forced to work part-time, or under employed. First, consider the people who are not in the labor force, but still want a job (row #11). The BLS reports this number every month, but I rarely see it in the media. This pool of people has grown from 2.0% of the civilian non-institutional population in 2007 to 2.8% in 2013. As you can see in the table, the broad definition of unemployment (U-6) only counts the 2.3 million people who the BLS defines as “marginally attached” (row #13) even though there are 3.9 million more who also want a job (row #12). These are the people who are relatively invisible. Thus, for my definition of people at risk, I count the entire 6.3 million who are out of the labor force, but say that they want a job (row #11). Second, consider the people who are working full-time, but are struggling. In my last post I said that jobs are increasing at the low and high-end of the income scale, while middle-income jobs are declining. How can we estimate the number of people who have full-time jobs, but are unable to afford necessities, unable to raise a family, unable to buy health insurance, or are otherwise working full-time, but are at risk?
To begin answering this question, one can get data from BLS and estimate the number of workers within any given income range. Unfortunately, this is difficult to do using current employment statistics (CES), the data used to generate the monthly report on the employment situation (as discussed above). The reason is that the BLS estimates workers and wages by industry, not by occupation. So, the average wage for janitorial services (series CES6056172003) is the total payroll divided by the total employees, which includes janitors, managers, bookkeepers, and everyone else employed at the companies that make up this industry. Also, there might be people working as janitors who don’t work in this industry. Instead, it is necessary to look at the occupational employment statistics (OES), which the BLS publishes annually, and provides detailed wage data by occupation. In this data base, for example, I can find the occupation “Janitors and Cleaners, Except Maids and Housekeeping Cleaners” (Code 37-2011). Since the last report was in 2012, I had to extrapolate the trend out to August 2013. Even for past years, the number of workers in a specific wage range might be under counted: in some years, data for a specific profession may not be available; every year, hourly data is sometimes not provided for occupations that don’t work year round. Nevertheless, with careful analysis, I was able to estimate the number of workers in any wage range for any year.
As a side note, the BLS reported in August 2013 that they have an experimental program to combine the wage estimates from these two surveys. This will be a welcome upgrade, but In the meantime I joined the data myself (in the table above, all the rows come from CES data except rows #4, #5, and #7. I estimated row #7 based on OES data and computed the other two rows using this number plus CES data. I should also mention that the BLS discourages year by year comparisons of OES data. This is because occupations constantly change (some jobs are invented and others become obsolete). In my case, however, I am adding up people who earn within a given wage range across all occupations. Therefore, I can compare my counts year to year provided that I use constant dollars.
To estimate the number of full-time workers at risk, I counted all the workers earning below the federal 1968 minimum wage, $1.60 per hour, scaled for inflation (the BLS provides a handy calculator). This is a first approximation of at-risk workers; later I’ll discuss how to improve this calculation. I chose 1968 because that’s when the value of the minimum wage peaked; the value of today’s federal minimum wage ($7.25 per hour) is much less than the 1968 wage in today’s dollars ($10.75). The Economic Policy Institute argues that the declining value of the minimum wage is one of the forces driving inequality. Looking at the table above, I estimated the number of at risk full-time workers at about 33 million (row #7). I give the numerator and denominator for this “At Risk Labor” calculation in rows labeled “AR”. The numerator is the sum of rows #5, #8 and #11, and the rate is 36%. Note that the denominator is the sum of rows #2 and #11: those not in the labor force, but still want a job, are added to the total civilian labor force for this calculation.
As stated earlier, more than one-third of the civilian non-institutional population is at risk. Using this approximation of at risk workers, here’s the trend for the past ten years:
My estimate of the total at risk labor is almost surely low. The social security administration reports that 61 million people who reported federal income taxes in 2011 (about 40% of wage earners) earned less than $20,000 per year. Of course this statistic includes my son, who was under 16 and earned extra money over the summer. Nevertheless, this count is greater than my estimate of 27 million full-time at risk workers in 2011. The Economic Policy Institute reports that the median family budget area in the nation is Topeka, Kansas, where a family of four would need to earn $63,364 for an adequate, but modest, living. With both parents working full-time (2080 hours per year each), they would need to earn more than $15 per hour. Such a family would not be included in my count. At the same time, you might argue that not everyone earning below $10.75 an hour is at risk. Young people who work and live with their parents, for instance, might not be at risk. I agree, and I’m open to improving my method for a future post. In principle, the BLS data could be analyzed by region, age, and other demographics. Also, there are many budget calculators available, such as the Living Wage Calculator (MIT) or the Family Budget Calculator (EPI) that can be used to find the minimum wage necessary to meet basic living needs for every region of the country. Nevertheless, I’m confident of two things: 1) my estimate of full-time at risk workers is lower than the true number and 2) the number of people who are at risk is growing.
Let me put this another way. The percentage of the population that is struggling to meet basic needs has grown since 2007, even while the official measure of unemployment has dropped. In New York City, the homeless shelter population is at a record 50,000, and many of these people work full-time, sometimes holding multiple jobs. In fact, many of these people earn only around $5.00 per hour (less than required by law), and appeals to the Department of Labor have often gone unanswered. My own town, Bedford, MA, houses about 90 homeless families. Many of these people go to work every day, while their children go to school, and must make do with a single room equipped with only a microwave oven. We need a better way to measure the total at risk population. While it is true that jobless claims are at a six-year low, I disagree with economists, such as Paul Ashworth, concluding that this represents an “improvement in job market conditions.”
Employment Situation Summary
To be clear, there are opportunities for some, and this will continue. A newly updated paper by Emmanuel Saez, who is a Professor of Economics and, among other things, tracks the distribution of income, says that “the top 1% captured 95% of the income gains in the first three years of the recovery.” I recently had dinner with the Chief Technical Officer for one of my clients in the investment industry. He talked at length about the difficulty of hiring and keeping talented college graduates to work on “big data” and “analytics,” two of the hot areas these days. There is fierce competition for talent from large software companies as well as start-ups. Speaking of start-ups, I recently went to a product showcase where seven emerging companies had five minutes to show their product and five minutes for Q&A. What struck me was how easily a tiny group of people could quickly amass a huge user base. For example, one company had only four people and over 100,000 users. This is possible, in part, because many vendors can provision the digital infrastructure to run an online company in hours, with no capital investment.
Nevertheless, life is hard these days for many people in the United States. Consider these data points:
- The labor force participation rate is at its lowest level since 1978
- Two-fifths (37.9 percent) of the 11.3 million unemployed people in August 2013 had been looking for work for more than 6 months
- Once a person is unemployed for 6 months or longer, it becomes extremely hard to find employment
- In the US, the average household income is less than when the recession ended, falling 4.4% since June 2009
- Wages have been stagnant or declining for most people since 2000
- Fast food workers are walking off the job to demand higher pay
Thomas Friedman suggests that young people today will need to “invent” their job, and not “find” their job. He says “they will have to reinvent, re-engineer and re-imagine that job much more often than their parents.” Friedman’s advice has clearly worked for some people, such as George Popescu, the young CEO of Boston Technologies who has three master’s degrees (one from MIT). The problem is that not everyone will be able to become an innovator; not everyone will be able to invent their job. We can’t stop the coming changes to our economy, but we do need to make sure that our future economy supports those with low to medium skills so that they too can make a meaningful contribution to society. All of us have family, friends, and neighbors who need to support themselves and can’t do more.
This is post three of five. Next week, I’ll talk about the digital revolution and the accelerating pace of technological change. I’ll return to talking about the people I introduced in my last post; people who are leading the way to explaining how this story will unfold as known economic forces mix with the changing landscape of technology and automation, including robotics. In my final post of this series, I’ll talk about how these forces will come together in the next decade to create a perfect storm. Although I remain optimistic about our long-term future, I also believe getting there will be a bumpy ride.