John Schmitt discusses \”Low-wage Lessons\” in a January 2012 paper written for the Center for Economic and Policy Research.

Define \”low-wage jobs\” as those that involve earning two-thirds or less of the median hourly wage: that is, those earning less than about $10/hour. As Schmitt notes: \”If low-wage work were a short-term state that helped connect labor-market entrants or re-entrants to longer-term, well-paid employment, high shares of low-wage work would be less of a social concern. Indeed, if low-wage work facilitated transitions from unemployment to well-paid jobs, countries might want to encourage the creation of a low-wage sector to improve workers’ welfare in the long term.\” On the other side, if low-wage jobs are a near-permanent state of affairs for a substantial group of workers, or if such jobs even send a negative signal to potential future employers that this worker is going to have low productivity, then the prevalence of low-wage jobs may be of real policy concern.

Given the rising levels of inequality in the U.S. economy in recent decades, it\’s not a big surprise that the share of workers who can be classified as \”low-wage\” has been rising, from about 22% of the workforce in 1979 to about 28% of the workforce by 2009.

Moreover, the share of U.S. workers who are low-wage is considerably higher than in many other high-income countries. About one-quarter of U.S. workers are low-wage, compared with 20-21% in the UK, Canada and Germany; about 15% in Japan; and 8% in Norway and Italy.

The issue here can be summed up with this question: If someone in the U.S. economy is a law-abiding citizen who works full-time for a period of years, can they earn a level of wages that let them afford a slice of middle-class standard of living? If you are earning $10/hour and working 2,000 hours per year, your annual earnings of $20,000 would put you below the poverty line of $22,891 for a single parent with three children. 

And the problems of low-wage work aren\’t limited to low wages. Schmitt writes: \”Not only are low-wage workers likely to stay in low-wage jobs from one year to the next, they are also more likely than workers in higher-wage jobs to fall into unemployment or to leave the labor force altogether. … U.S. labor law offers workers remarkably few protections. U.S. workers, for example, have the lowest level of employment security in the OECD and no legal right to paid vacations, paid sick days, or paid parental leave. … [M]ore than half (54 percent) of workers in the bottom wage quintile did not have employer-provided health insurance and more than one-third (37 percent) had no health insurance of any kind, private or public.\”

It\’s worth noting that labor force participation rates for men aged 16-24 have fallen from 72% in 1990 to 57% in 2010, and for men from 25-54, the labor force participation rate has fallen from 93% in 1990 to 89% in 2010, according to Bureau of Labor Statistic data.  Much of this is due to the low pay available to those with low skill levels.

Schmitt only sketches his policy suggestions here, which include higher rates of unionization, higher minimum wages, employment-protection legislation and other national labor laws, along with higher benefits for the jobless and low-income households. He less of a fan of the Earned Income Tax Credit, fearing that employers capture much of the benefit of the credit because it allows them to pay lower wages than they otherwise would. For my own part, dramatically higher rates of unionization would fly in the face of a half-century trend in the U.S. (see this post for some details).  While I\’m comfortable with the minimum wage playing some role in the labor market, jacking it up by 50% or more seems to me unwise.  I\’m an enthusiastic supporter of the EITC, and a cautious supporter of certain national legislation to improve employment benefits and conditions.

But my purpose here is not to argue policy, but only to point out that the U.S. labor market seems to be producing an outcome where a substantial and growing proportion of full-time employees earn barely enough to creep above the poverty line. If we wish to build a society and an economy on rewarding work, it is a harsh fact of U.S. labor markets that such a reward is currently not apparent for many.

Margaret Jacobson and Filippo Occhino of the Cleveland Fed offer a short overview of what is \”Behind the Decline in Labor\’s Share of Income\” in the February 2012 issue of Economic Trends.

Back in the day, when I was first getting familiar with these numbers, the standard summary of the data was that labor income was about two-thirds of total output in the U.S. economy, although the share fluctuated over the business cycle. As Jacobson and Occhino write: \”Over the cycle, the labor income share tends to increase during the early part of recessions, because businesses lower labor compensation less than output, and compensation per hour continues to increase even as productivity slows down. Then, after reaching a peak sometime during the recession, the labor income share tends to decrease during the rest of the recession and the early part of the recovery, as output picks up at a faster pace than labor compensation, and compensation per hour grows at a slower pace than productivity. Only later in the recovery, as the labor market tightens, does labor compensation catch up with output and productivity, and the labor income share recovers.\”

But this basic fact–labor as two-thirds of economic output–no longer seems to be holding true. It\’s not just that the ratio is at historic lows in the post-World War II period, as shown in the figure; after all, given the depth and length of the Great Recession, and the sluggishness and sustained high unemployment of the tepid recovery, it\’s no surprise that the labor share of income would be low about now. But the data seems to show an overall pattern of a dropping labor share of income even before the Great Recession started, and reaching back to the 1980s or the late 1970s.

When output is rising faster than labor income, it necessarily follows that labor compensation is rising more slowly that output per hour. Here\’s the figure. Note that up until about the early 1980s, productivity as measured by output per hour rose more-or-less in step with compensation per hour (although it appears that even then, output/hour was trying to creep ahead). But the gap has expanded since then, and was expanding even before the Great Recession.

What explains this change? Jacobson and Occhino list the possibilities: \”Economists have identified three long-term factors that can explain why the wage-productivity gap has widened and the share of income accruing to labor has declined. The first is the decrease in the bargaining power of labor, due to changing labor market policies and a decline of the more unionized sectors. Another factor is increased globalization and trade openness, with the resulting migration of relatively more labor-intensive sectors from advanced economies to emerging economies. As a consequence, the sectors remaining in the advanced economies are relatively less labor-intensive, and the average share of labor income is lower. The third factor is technological change connected with improvements in information and communication technologies, which has raised the marginal productivity and return to capital relative to labor.\”

This list seems basically right to me, although my reading of the evidence is that the items are listed in inverse order of importance. But I also find it useful to think about the fall in labor income in reverse, as the rise of capital income. For example, the Dow Jones index roughly tripled from 1980 to 1990 (rising from 900 to 2700), and then more than tripled from 1990 to 2000 (rising from 2700 to 10,800). While the Dow was basically flat over the decade from 2000 to 2010, large non-labor income was generated during the earlier part of the decade by housing prices. While many of us participate in gains in the stock market or the housing market in some ways, the bulk of those gains do tend to flow to those with higher income levels.

But looking ahead, another rapid tripling of stock market, as in the 1980s and again in the 1990s, seems unlikely, as does another housing price bubble. Meanwhile, the U.S. labor market is showing some feeble signs of resurgence, and may well strengthen over the next couple of years. Over the next 3-5 years, I\’d expect the labor income share of output to rise–although I don\’t expect labor income to re-attain the old level of two-thirds of national output in the near- or the middle-term.

Government policies toward unemployment fall into two categories: passive and active. Passive policies are those like unemployment benefits or early retirement. They tide over the affected workers until the next job, or until retirement, but accomplish little else. Active policies are those like government support for job training, job search, incentives for private firms to hire, or even direct job creation. The U.S. spends less of either kinds of labor market policy than most developed economies. Here, I\’ll draw on a recent essay by Jun Nie and Ethan Struby from the Kansas City Fed: \”Would Active Labor Market Policies Help Combat High U.S. Unemployment?\”

Let\’s set the stage with some basic facts about much developed economies typically spend on active and passive labor market policies. Here\’s a bar graph from Nie and Struby showing the average levels over the 1998-2008 period for both categories of labor market spending for 21 countries.

The United States spends less on labor market policies overall than most developed economies, and of what it spends, a smaller proportion goes to active labor market policies. Nie and Struby describe the patterns this way: \”The level of spending on labor market policies differs widely across OECD countries. Between 1998 and 2008 in 21 OECD countries, total expenditures on passive and active labor market policies as a fraction of GDP ranged from about 4 percent in Denmark to 0.25 percent in the United Kingdom (Chart 3). The United States is near the bottom of this list, spending slightly less than 0.5 percent of GDP on labor market policies during this time. In addition, the fraction of spending on active versus passive policies differs across countries. Outside the United States, the average country in Chart 3 devoted 59 percent of labor market policy expenditures to PLMP [passive labor market policies] and 41 percent to ALMP [active labor market policies]. In the United States, however, 70 percent of expenditures went to PLMP and 30 percent went to ALMP.\”

Nie and Struby offer an alternative metric: Instead of looking at spending as a share of GDP, look at expenditures per unemployed worker. Moreover, measure those expenditures as a percentage of per capita GDP. By this measure, U.S. spending on passive labor market policies is half as large as the average for other developed economies, while U.S. spending on active labor market parties is a bit more than a quarter of the average for other developed economies. They write:

\”Another common method to measure spending on labor market policies is to consider expenditures per unemployed worker as a percentage of GDP per capita. This measure adjusts for differences across countries in unemployment rates and the size of the economy. Between 1998 and 2008, average expenditures on PLMP in the United States on each unemployed worker were about 12 percent of GDP per capita, while the average level for the 20 other OECD countries in Chart 3 was about 25 percent of GDP per capita. The U.S. expenditures on ALMP were even less: Over the same period, expenditures on ALMP per unemployed worker in the United States were about 5 percent of GDP per capita, while the average spending per unemployed worker for the 20 other OECD countries was approximately 19 percent of GDP per capita.\”

How well do active labor market policies work? Nie and Struby offer a reasonable first take on this question, but it should be noted that here we are departing the world of basic facts about categories of spending levels, and entering the world of estimation. Details of the their calculations are in the article. Here, I\’ll just say that they look at the spending by these 21 countries with active and passive labor policies over the 1998-2008 period, and they break down these policies into various subcategories. They look at how unemployment rates fluctuate in those countries. They also adjust for a bunch of other variables: the labor force participation rate, union density, laws about employment protection, tax rates, whether the economy is in recession, level and duration of unemployment benefits, and \”fixed effects\” that are supposed to account for country-specific factors like cultural, geographical. and political differences.

Again, this is all a reasonable starting point, but as the authors are careful to note, the the \”results should be read cautiously.\” This calculation has lots of correlations, but causation is less clear. Extrapolating from average experience in some countries at one time to a particular country at another time is always somewhat questionable.

With warning flags duly posted, the results of their calculations are intriguing. They find that among the active labor market policies, expansions in job training opportunities and assistance with job search are the least expensive ways to encourage more employment, and that these methods pass a rudimentary cost-benefit test. They also find that employment incentives and direct government support for employment are much costlier and under at least some sets of plausible assumptions, don\’t pass a cost-benefit test.

To me, this all adds up to a tentative but plausible case for more government spending on job training programs and on support for job search. The U.S. traditionally hasn\’t done much for its unemployed–neither passive nor active support–as compared with other countries. But with U.S. unemployment high, in an economy where lifelong learning and movement between economic opportunities is ever more important, these seem like potentially useful steps.A focused program of widespread state-level and local-level experimentation with these policies, together with studying the experience of other countries, could lead to policies that are useful for U.S. labor markets both in the next few years and in the long-run.

Finally, although the authors don\’t make this point, the two active labor market policies that they find useful–assistance for job training and for job search–both focus directly on the unemployed. The two programs they don\’t find useful–employment incentives and government support for employment–both focus on the potential employers, and seem much more susceptible to gaming the system, and to politicized and pork-barrel spending.

The Bureau of Labor Statistics has just put out its \”Job Openings and Labor Turnover Survey Highlights\” with data up through September 2011–and lots of nice graphs and explanations. The 2010 Nobel Prize in Economics went toPeter A. Diamond, Dale T. Mortensen, Christopher A. Pissarides \”for their analysis of markets with search frictions.\” Their work is a reminder that unemployment is not just about a shortfall in demand, but is also a matter of search and matching by potential workers and employers. The JOLTS data offers the factual background on job openings, separations, hires, and more. The overall picture is of an unpleasantly stagnant labor market.

As a starting point, look at the relationship between hires, separations, and employment. Most of the time, the red line showing job separations and the blue line showing hires are pretty close together, with one just a bit over the other. When separations exceed hires for some months running in the early 2000s, total employment declines. Then hires exceed separations by a bit, month by month, and total employment grows. During the Great Recession, hiring drops off a cliff and separations fall sharply as well (more on that in a second). Total employment has rebounded a bit since late 2009, but it\’s interesting to note that the levels of hires and separations remain so low. Those with jobs are tending to stay in them; those without jobs aren\’t getting hired at a rapid rate.

What explains why job separations would fall during the Great Recession? After all, don\’t more people lose their jobs in a recession? Yes, but the category of \”job separations\” has two parts: voluntary quits and layoffs/discharges. During the recession, layoffs and discharges do rise sharply as shown by the red line, but quits fall even faster as shown by the blue line, as those with jobs hung on to them. Overall, job separations decreased. Notice that in the last year or so, layoffs and discharges have actually been relatively low compared to the pre-recession years in the mid-2000s. Quits have stayed lower, too.

The number of job openings in an economy tends to be a leading indicator for changes in employment. Notice the sharp drop in job openings in the recession of 2001, and again the sharp drop in job openings during the 2007-2009 recession. Employment levels decline soon after. However, it\’s interesting to note the upturn in  job openings since the low point in July 2009, and how employment has correspondingly grown.

Looking at the ratio of job openings to the unemployed gives a sense of how difficult it is to find a job at any given time. Before the recession of 2001, there were about 1.2 unemployed people per job opening. In the aftermath of the \”jobless recovery\” from that recession, there were about 2.8 unemployed people per job opening in late 2003. There were about 1.5 unemployed people per job opening in mid-2007, but just after the end of the recession in late 2009, there were almost 7 unemployed people for every job opening. This statistic helps to emphasize that it isn\’t just that the unemployment rate remains high, but that unemployed people in a stagnant labor market, with low hiring and few separations, objectively will have a hard time finding jobs.

A final figure from this data is called the Beveridge curve. BLS explains: \”This graphical representation of the relationship between the unemployment rate and the vacancy rate is known as the Beveridge Curve, named after the British economist William Henry Beveridge (1879-1963). The economy’s position on the downward sloping Beveridge Curve reflects the state of the business cycle. During an expansion, the unemployment rate is low and the vacancy rate is high. During a contraction, the unemployment rate is high and the vacancy rate is low.\” The figure is usefully colored in time segments, so the period before the 2001 recession is in light blue in the upper left corner; the 2001 recession is the darker blue line t; the period of growth in the mid-2000s is the red line; the Great Recession is the green line; and the period since the recession is the purple line at the bottom right. The severity of the Great Recession is apparent as the green line stretches down to the right, with much higher unemployment and lower rates of job openings than the 2001 recession.

But the Beveridge curve also raises an interesting question: Is the economy getting worse at matching people with jobs? The low levels of  hiring and separations suggest a stagnant labor market. The Beveridge curve might be another signal. As BLS explains: \”The position of the curve is determined by the efficiency of the labor market. For example, a greater mismatch between available jobs and the unemployed in terms of skills or location would cause the curve to shift outward, up and toward the right.\” Notice that as the number of job vacancies has increased, the unemployment rate hasn\’t fallen as quickly as one might expect. To put it another way, the purple line is not retracing its way back up the green line of the Great Recession, but instead is above it and to the right.

Of course, relationships in the economy aren\’t going to be as precise as, say, the relationship between altitude and air pressure. There isn\’t yet enough data to prove whether the Beveridge curve has in fact shifted out. But if it has indeed become harder in the U.S. economy to match unemployed workers with job openings–perhaps because the skills that employers are searching for are not the same as the skills that the unemployed have to offer?–then it will be even harder to bring down the unemployment rate.  

When information is imperfect, markets may not work well. If consumers are highly uncertain about the quality of what they are buying, they become less likely to buy. If a lender is highly uncertain about whether a potential borrower will repay a loan, the lender is less likely to make that loan. The more uncertain that an employer is about the quality of a potential employee, the less likely the employer is to hire that person. This problem of imperfect information in labor markets is especially severe now, with an unemployment rate that has been between 8.8% and 10.1% since April 2009. If someone hasn\’t been working, how can an employer judge their skills and talents?

Is there a way in which workers with experience could have a way of demonstrating their skills and knowledge that didn\’t involve taking a class or getting a degree? There are some recent experiments along these lines in the U.S. economy, but it turns out that several other countries have already developed processes for recognizing  non-formal and informal learning.

Jeff Selingo, who is editorial director of the Chronicle of Higher Education, tackled the question of whether colleges might lose their near-monopoly power over anointing people with job credentials in a short article last month. Selingo writes:

The day when other organizations besides colleges provide a nondegree credential to signify learning might not be as far off as we think. One interesting project on this front is an effort to create “digital badges,” which would allow people to demonstrate their skills and knowledge to prospective employers without necessarily having a degree. Badges could recognize, for example, informal learning that happens outside the classroom; “soft skills,” such as critical thinking and communication; and new literacies, such as aggregating information from various sources and judging its quality. And in a digital age, the badge could include links back to documents and other artifacts demonstrating the work that led to earning the stamp of approval.

Until now an interesting-but-somewhat-fringe idea, digital badges received a big boost last week, when the John D. and Catherine T. MacArthur Foundation announced a $2-million competition to create and develop badges and a badge system. (The contest is also supported by Mozilla and the Humanities, Arts, Sciences, and Technology Advance Collaboratory, otherwise known as Hastac.)

At the announcement in Washington, the U.S. secretary of education, Arne Duncan, called badges a “game-changing strategy” and said his agency would join with the Department of Veterans Affairs to award $25,000 for the best badge prototype that serves veterans looking for well-paying jobs. Under a badge system, colleges would no longer be the sole providers of a credential. While badges could be awarded by traditional colleges, they could also be given out by professional organizations, online and open-courseware providers, companies, or community groups.

In the Autumn 2011 issue of the Wilson Quarterly (not available free online), Kevin Carey writes in an essay about \”College for All?\” about how the Western Governors University is awarding degrees based on competency, not classroom hours. Carey writes:

\”While American higher education is diverse in many ways, encompassing a variety of missions and constituencies, it is remarkably undiverse when it comes to awarding degrees. Every institution grants the same two- and four-year credentials that signify little more than how many hours the bearer sat in classrooms. Newer institutions such as Western Governors University (WGU) are turning that equation upside-down, awarding degrees when students demonstrate defined competencies, regardless of how long it took to achieve them.\”

WGU is a fully accredited nonprofit institution founded in the 1990s by the governors of 19 western states that now enrolls 25,000 mostly adult students online. It currently focuses on occupation-specific fields such as education, business, and health care. But efforts are afoot to expand the model into more traditional academic fields.

The WGU experiment points to a future public education system in which public subsidies are tied to commonly understood goals for learning, not how old the student happens to be or whether he or she happens to live. In increasingly digital learning environments, it will be possible to track, store, and summarize evidence of learning in ways that render traditional time-based credentials obsolete.\”

On the international front, Patrick Werquin wote an OECD report on \”Recognising Non-Formal and Informal Learning: Outcomes, Policies and Practices\” which was published in spring 2010 (and to my knowledge is not freely available on-line). Some highlights (omitting some references for readability): 

\”All data on lifelong learning indicate that the highest qualification held by the great majority of people is obtained in the formal system of education and initial training, which in the case of many adults occurred some time ago. This is confirmed by other sources revealing that almost 90%of adult learning initiatives do not lead to a qualification, even though, depending on the country, 20-60% of  individuals who embark on learning do so primarily to obtain one. … There is therefore a patent lack of visibility as regards people\’s real knowledge, skills and competences, since those acquired during their working lives or other activities remain invisible. This lack of visibility is all the more significant for those who left the initial education and training system many ears earlier. It is also especially detrimental to those with a low level of qualification …\”

\”More recently, OECD (2007) ranked the recognition of non-formal and informal learning outcomes high on a list of 20 mechanisms identified as potentially capable of motivating learning. At the same time, major international organizations are showing a close interest in the recognition of learning outcomes. All these studies point in the same direction: formal learning alone cannot account for all of the learning encompassed by the concept of lifelong learning. There is thus no shortage of studies that argue for the recognition of non-formal and informal learning outcomes. …\”

Werquin\’s report for the OECD lists mechanisms for recognizing non-formal and informal learning in 21 countries–notably, with no mention of any such effort in United States. Two countries with especially well-developed policies along these lines are Ireland, which has certificates for Recognition of Prior Learning (RPL), Accreditation of Prior Experiential Learning (APEL), Recognition of Current Competences (RCC), Learning Outside Formal Teaching (LOFT) and others, and Norway, which has a \”skills passport\” system.

These systems for recognizing non-formal and informal learning systems vary considerably across countries. I can imagine a number of practical concerns. But for many Americans, maybe especially those with lower and medium skill levels, their educational credentials (often from long-ago) don\’t reveal their true skill set. For many of them, going back to school for some additional degree or certificate is impractical, and frankly a waste of time–because whether they have a piece of paper from an educational institution to prove it, they have already acquired the skills they need for many jobs. America should be thinking more about ways of connecting potential workers to the labor market that don\’t involve telling those who don\’t flourish in school that they need to keep attending. A couple of weeks ago I posted on Apprenticeships for the U.S. Economy as one such option. Ways to recognize competences achieved through non-formal and informal learning seems like a complementary approach.