Between June and July of this year, 44,000 nonfarm jobs disappeared, according to the Labor Department’s Bureau of Labor Statistics.

Between June and July of this year, 44,000 nonfarm jobs disappeared, according to the Labor Department’s Bureau of Labor Statistics. But during the same period, BLS reported a 0.2 percent drop in the unemployment rate. Why didn’t unemployment go up instead of down? Here’s another conundrum: In 2002, the American economy, as measured by real gross domestic product, grew 2.4 percent, yet the number of Americans employed dropped 222,000. Is such a “jobless recovery” really possible?

Data on trends in employment and unemployment are vital to judging how well the nation’s economy is performing, but figuring out what those data mean is complicated. The federal government reports employment and economic performance data in several different ways, which sometimes leaves people more confused than edified. In this issue of IRConcepts, we attempt to solve the riddle of the missing jobs by examining the definitions of employment and unemployment and the varying ways of measuring them. Along the way, we explore the workings of the labor market, and see how they differ in practice from the neoclassical economic model. By the time we’re done, we should know why those 44,000 jobs that disappeared in June did not show up in the unemployment data and how an economy can bounce back from recession without creating new jobs. We should also be able to make some predictions about the future employment picture. First, however, we need to understand some basic labor economic terms and concepts.

A market is a place where goods and services are exchanged through the medium of money. Like any other market unencumbered with artificial constraints, a free labor market theoretically would “clear,” that is, maintain a balance of supply and demand. In a free labor market workers would not be bound to an employer and could change jobs whenever they wanted. Employers, in turn, would be free to hire whomever they pleased, would have no obligations to their employees other than paying them for the work performed at the competitive price dictated by the market, and could dismiss them whenever their services were no longer required. Over time, custom, employer practice, unionization, and government regulation have modified this laissez-faire approach in capitalist countries. (More in some than in others—see the table below for a contrast of the European and the American labor market models.)

There is no single European labor market model, but certain features are common to many countries. The table below compares the typical European system with that of the United States.

Critics of the U.S. model claim that it gives employers too much latitude, leading to a lack of job security, a large number of low-wage jobs, and increasing inequality. Its great success, however, has been in job creation. The European labor market, on the other hand, is accused of inflexibility resulting in high unemployment and a dearth of job creation. The International Monetary Fund has even questioned whether European labor market regulations and institutions actually reduce inequalities:

European Labor Market Model

Strong union influence in government through social democrat or labor parties

High union representation

Centralized labor market and collective bargaining

Legislation restricting employers’ ability to lay off workers

High minimum wages

U.S. Labor Market Model

Labor unions influence government as do other interest groups, but not directly linked to a party and weaker than in Europe

Low union representation

Decentralized collective bargaining
Generous government-paid social benefits protecting workers from vicissitudes of market economy

Less generous safety net provided by unemployment insurance and social security

Minimal government regulation regarding employers’ macro employment decisions

Low minimum wage relative to average earnings

“To the extent that these policies primarily protect those already employed and have the perverse effect of increasing unemployment—which tends to be concentrated on the most vulnerable and least skilled segments of society, often including first-time job seekers—they may increase rather than reduce inequality.” (World Economic Outlook, May 1994)

Indeed, a number of countries, e.g., the Netherlands, have moved to more flexible labor markets in recent years. At the moment, both the French and German governments are engaged in political battles to enact modest reforms.

The labor market is peculiar in that it involves human beings, who cannot be treated like other commodities that are bought and sold. The public rarely notices when a company reduces the number of nuts and bolts it uses, but they certainly do when it cuts the number of its employees. Once a company purchases a dozen nuts and bolts it has full control over them, but an employee can decide to work more or less diligently or to quit. Furthermore, nuts and bolts of a given size are pretty much the same; one can be substituted for another. Not so with employees, particularly those who are more skilled than others. In addition, institutional arrangements may restrict what the employer can do—the government may legislate a minimum wage, and workers may form unions through which to bargain over wages, hours, and working conditions.

Even absent institutional interference, the labor market has not behaved strictly according to theory. Neoclassical economics posits that if for some reason the current supply of labor is greater than the amount demanded by employers, some workers will withdraw their services while others will lower the wage they are seeking, inducing employers to increase the amount of labor they are willing to hire. Any unemployment that would persist, therefore, would be considered voluntary, that is, due to workers’ choice to refuse the wages being offered. This view became untenable during the Great Depression of the 1930s, when people were willing to work for as little as a sandwich and a cup of coffee but still could not find jobs. This is because the demand for labor is a derived demand; it depends on demand for the goods and services that labor helps to produce. If for some reason the total of consumer, business, and government purchases are depressed, workers are not needed, regardless of how little they are willing to accept in pay.

The labor market is dynamic, with people continually flowing in and out as changes in demand and supply of goods and services affect the demand for labor overall, as well as for particular types of workers. In the late 1990s, the labor market was “tight” and employers scrounged to find all sorts of workers. In today’s loose market employers have little difficulty in hiring. At other times the situation has been mixed, as was the case a decade ago when there were plenty of unskilled workers but a dearth of those more skilled, such as computer scientists and systems analysts.

The demand for labor can change more quickly than the supply. While it is true that employers may attract a few more people into the labor force with offers of higher pay and amenities or may convince some to leave through inducements to retire early, basically the labor supply is fixed in the short run. Long-term shifts in the labor force depend on birth rates, immigration, and societal attitudes, which change slowly.

So far we’ve been using the terms labor force (and its rough synonym, labor supply) pretty freely without defining them precisely. In its reports BLS uses “labor force” in a very specific way to mean all civilians who are not institutionalized (e.g., in prisons or hospitals) and who seek and are capable of performing work, for themselves or for others, for remuneration. The definition therefore takes into account not only the size of the population of working age (16 years or older), but the decisions those individuals make about whether or not to work or pursue work.

To determine the size of the labor force, every month the Bureau of the Census conducts a current population survey for BLS. The survey covers some 60,000 households (so is also called the household survey), and assigns appropriate weights to the sample by location, race, age, sex, industry, and occupation. Respondents are classified as either in the labor force or not, depending on how they answer survey questions about their employment and/or job search activities. Those classified as not in the labor force consist primarily of homemakers, retired persons, and students, but, as we shall see, also include other jobless people who do not meet the precise criteria for inclusion.

The unemployment rate is the percentage of that labor force that cannot find jobs. In 2002 there were, on average, 217,570,000 Americans 16 years old and over who were not institutionalized or serving in the military. Of these, 66.6 percent (144,863,000 individuals) were actually in the labor force, that is, they were employed or were actively seeking work. The average number of Americans in the labor force without jobs was 8,376,000, yielding an unemployment rate for the year of 5.8 percent.

As intimated above, the way employment and unemployment are defined by the household survey has a big impact on the numbers reported. The definition of employment is broad; people are counted as employed if they did any work at all for pay or profit during the survey week or if they worked 15 hours or more without pay in a family-owned enterprise. Everyone engaged in part-time and temporary work, as well as regular full-time year-round employment, is counted as employed. Also counted are people who have a job at which they did not work during that week, because of a vacation, maternity/paternity leave, a labor dispute, or bad weather.

The definition of unemployment (i.e., those in the labor force but without jobs) is more restrictive, omitting many people who would like to work. Persons are classified as unemployed if they do not have a job, have actively looked for work in the prior four weeks, and are currently available for work. Active job search requires such steps as going on job interviews, sending out résumés, answering want ads, and checking union or professional registers. A jobless person who had merely searched the classifieds would not be considered unemployed—or even a member of the labor force. On the other hand, workers expecting to be recalled from layoff are counted as unemployed, even if they have not engaged in active job search.

The household survey defines many jobless people as not being in the labor force rather than unemployed, because they have not met any of its criteria of active job search. Some jobs losers do, in fact, exit the labor force, either permanently (e.g., older workers who retire) or temporarily. Job losers who seek work but are unable to find it may eventually give up looking; if so, they are not longer counted as unemployed. From June to July 2003, for example, the labor force was reported to have declined by 556,000, most likely because of these hidden unemployed who are no longer counted as part of the labor force.1 Unemployment also is held down by the fact that people who had been planning to seek work postpone doing so. Young adults and teenagers may postpone entering the labor force; from 2000 to 2002, for example, the teenage (16 to 19) labor force participation rate dropped from 52 to 47.4 percent. Women who had left the labor force temporarily to care for young children may postpone reentering. Evidence of this can be seen in the female labor force participation rate, which had been rising steadily for decades; it peaked at 63 percent in 2000 and was down to 61.1 percent in 2002. Putting off entering or entering the labor market when jobs are scarce also helps to explain why on the economic upswing unemployment does not drop as rapidly as jobs increase—the postponers come rushing back in search of work.

There are other ways of studying unemployment trends besides counting those who meet the household survey definition. For example, state unemployment insurance (UI) programs report the number of people who qualify for unemployment insurance. However, most jobless workers normally do not receive UI benefits, so these figures cannot be used as a count of total unemployment in the nation. Also, the number of unemployed who receive benefits varies depending upon whether extended benefits are available for people who have otherwise exhausted their unemployment eligibility. In 1988, when there were virtually no extended benefits, only 31 percent of the unemployed received UI payments, but when extended benefits were in effect in 1992, 51 percent did.

Despite their shortcomings, the UI data offer clues to market improvement or deterioration. Economists look at the four-week moving average of new claims as a fairly reliable indicator of trends. Any figure above 400,000 is considered to be a sign of labor market weakness. For the week ending July 19, 2003, 419,250 new claims were filed. The number of people who have exhausted their benefits is another important figure, as is the total number of people receiving benefits, excluding new claimants.

Further confusion regarding what is happening with respect to employment arises from the fact that, in addition to the household survey, there is another employment-related survey conducted every month. The establishment survey counts the number of people on the payroll in a sample of 160,000 nonfarm firms and government employers, including 400,000 separate establishments such as factories, offices, and stores and federal, state, and local government agencies. (Employees in these establishments are any workers who received pay for any part of the reference period.) By showing the number of total employees and production workers by industry, their hours worked, and their pay, the establishment survey pinpoints the sectors of the economy in which jobs are growing or declining.

The household survey and the establishment survey often seem to report contradictory numbers.2 For example, according to the household survey 137,430,000 people were employed in the first quarter of 2003, but the establishment survey reported only 130,225,000 on payrolls. The apparent discrepancy arises from conceptual and methodological differences between the surveys. The household survey includes agricultural workers, the self-employed, unpaid family workers, and private household workers among the employed, but the establishment one does not. On the other hand, a person with two jobs is counted twice in the establishment survey, but only once in the household one.

Although most attention is focused on unemployment, more significant insight into the state of the labor market may be provided by the employment ratio—the percentage of the noninstitutional population, age 16 and over, actually working. As the labor force and the economy have grown, so has the employment/population ratio. In 2000, it reached 64.4 percent, the highest in the nation’s history. By 2002, however, it had declined sharply to 62.7 percent and has continued to decline in 2003, a sign that the labor market is deteriorating seriously and that the nation’s economy is operating well below its potential.

Often charges are leveled against the official count claiming that much unemployment is “hidden.” Disagreement over the accuracy of the data stems from the fact that people want to use them for different purposes. Thus, while the data may be adequate to determine how well the economy is utilizing the available supply of labor, they may not be sufficient as a measure of the economy’s performance in relieving hardship for jobless people, since many of these are simply not counted as unemployed. BLS is not unmindful of the problem. Included in the monthly household survey are questions for those classified as not in the labor force regarding their desire for work, why they have not sought work in the last four weeks, their prior job search, and their availability for work. Their answers form the basis for estimating the number who are not in the labor force but are “marginally attached” to it (1.5 million in June 2003). Such persons, though currently not looking for work, want to work and are available. Discouraged workers, a subset of the marginally attached group, are those not currently looking for work because they believe none is available in their line of work or locale or because they previously had not been able to find work, they lack education and training required, or employers consider them too young or too old.

To get a fuller picture of unemployment, in 1976 Julius Shishkin, then commissioner of labor statistics, suggested a number of additional measures of unemployment. BLS reports six, labeled U-1 through U-6 in the list below. (The first three are seasonally adjusted; the others are not.)

Code DefinitionJune 2003 Rate
U-1 Persons unemployed 15 weeks or longer, as a percentage of the civilian labor force (the rationale is that unemployment of such length causes economic hardship)2.4%
U-2 Job losers and persons who completed temporary jobs, as a percentage of the civilian labor force 3.4
U-3 Total unemployed, as a percentage of the civilian labor force (official unemployment rate) 6.4
U-4 Total unemployed plus discouraged workers, as a percentage of the civilian labor force, plus discouraged workers 6.8
U-5 Total unemployed plus discouraged workers plus all other marginally attached workers, as a percentage of the civilian labor force, plus all marginally attached workers 7.4
U-6 Total unemployed plus all marginally attached workers plus total employed part-time for economic reasons, as a percentage of the civilian labor force, plus all marginally attached workers 10.6

These data give us a more complete picture of the labor force but still do not tell us about the impact of unemployment on individuals. Although unemployment in 2002 averaged 8,378,000, this does not mean that 8.4 million specific individuals were unemployed during the course of the year. Some of those who were jobless one month might have been back working the next, while others who had been laid off or entered the labor market took their places. As a result, during the course of the year many more millions suffered some period of joblessness.3 The 2002 figure is not available yet, but in 2001, when unemployment averaged 6.8 million, 15.8 million people experienced some unemployment, and applying that ratio to 2002 translates into 20 million being unemployed for some period during the year.

In good times, most of the people who are unemployed have not lost jobs, but are new entrants, reentrants, and voluntary quits. Jobs are available, but it takes a few weeks for a person to connect with a suitable one. The situation reverses dramatically as the labor market deteriorates. Employers lay off workers, those with jobs tend to hang on to them, and many not in the labor force postpone entering it. The data for June of this year illustrate the pattern: 5 million people, or 54.2 percent of the total unemployed, lost jobs during the month. Another 0.9 million, or 9.7 percent, left their jobs voluntarily; 2.7 million, 29.1 percent, returned to the labor force; and only 0.6 million, 7 percent, were new to the labor market.

Not all workers are equally vulnerable to being unemployed. Summers found that most unemployment, even in tight labor markets, is attributable to relatively few persons who are out of work a large part of the time. There are marked differences in unemployment according to race, age, sex, education, industry, and occupation.

Blacks and Hispanics have much higher rates of unemployment than do whites. Regardless of the national level of unemployment, the black rate tends to be double that of whites and the Hispanic rate about 50 percent higher. The black-white ratio has remained virtually constant for decades. In 1972, when the overall unemployment rate was 5.6 percent, that of whites was 5.1 and of blacks, 10.4 percent; 30 years later, in 2002, when the overall rate was 5.8 percent, that of whites was 5.1 and of blacks, 10.2 percent. The explanation lies in the fact that blacks remain in the age groups, occupations, industries, and educational attainment brackets that are subject to greater unemployment, and some discrimination continues to exist. Forty years after enactment of civil rights legislation, there is still no level playing field. Blacks have made progress, but as a group they are still less skilled and less educated than whites, and the gap in unemployment rates between blacks and whites has not narrowed.

The current weak labor market has affected all demographic groups, as can be seen in Table 1, which depicts changes in employment and unemployment, by sex and race, from 2000 to 2002. The data reveal, however, that male unemployment increased more than female. Among males, whites suffered the greatest increase in unemployment, which probably stems from the fact that they are more highly concentrated in the hardest-hit sectors of the economy, such as manufacturing. Hispanics, a growing share of the labor force, actually had an increase in employment of 0.8 million, divided evenly between men and women, yet the men also suffered much higher unemployment. Asians, for whom there is no breakdown by sex, have had the sharpest jump in unemployment.4

Table 1: Employment and Unemployment by Race and Sex, 2000 and 2002 (U.S. Department of Labor; Bureau of Labor Statistics)

Employed Unemployed Rate Employed Unemployed Rate
Total 136,891 5,692 4.0% 136,485 8,378 5.8%
Men 73,305 2,975 3.9 72,903 4,597 5.9
Women 63,586 2,717 4.1 63,582 3,781 5.6
White 114,424 4,121 3.5 114,013 6,137 5.1
Men 62,289 2,177 3.4 61,849 3,459 5.3
Women 52,136 1,944 3.6 52,164 2,678 4.9
Black 15,156 1,241 7.6 14,872 1,693 10.2
Men 7,082 620 8.0 6,959 835 10.7
Women 8,073 621 7.1 7,914 858 9.8
Hispanic 15,735 954 5.7 16,590 1,353 7.5
Men 9,428 494 5.0 9,845 764 7.2
Women 6,307 460 6.8 6,744 590 8.0
Asian 6,043 227 3.6 6,215 389 5.9

Differences in unemployment rates are even sharper among age groups—the younger, the higher. In 2002, young workers (those under 25 years of age) made up 32 percent of all the unemployed, even though they were only 15 percent of the labor force. Among the young, females have lower rates than males. (See Table 2.) As unemployment rises, young workers suffer the most because as more mature workers become available, employers are less inclined to hire inexperienced youngsters.

Table 2: Rates of Unemployment by Age and Sex, 2002(U.S. Department of Labor, Bureau of Labor Statistics)
Age All Men Women
16 & over 5.8% 5.9% 5.6%
16-19 16.5 18.1 14.9
20-24 9.7 10.2 9.1
25-34 5.9 5.8 5.9
35-44 4.6 4.5 4.6
45-54 4.0 4.2 3.8
55-64 3.7 4.3 3.5
65 & over 3.6 3.4 3.9

Teen-age unemployment has been above 11 percent since 1954 and reached 23.2 percent in the recession year of 1982. (It must be recognized, however, that many of the teenagers are full-time students who only want part-time work.) Unemployment rates are much higher for black teenagers, averaging between 30 and 40 percent in many years. In June of this year, the teen-age rate of unemployment reached 19.3 percent—16.5 percent for whites and 39.3 percent for blacks. Black youth unemployment is concentrated in the poor areas of central cities, where the rate can approach 50 percent. And because more black than white youngsters are not in school, more want a full-time job. The racial disparities stem from differences in educational attainment, skills, unequal access to the informal networks through which people secure good jobs, and employer stereotyping of minority youth as high risk.

The type of job one has, in terms of both industry and occupation, also determines susceptibility to unemployment. Workers in construction and agriculture, in which work is seasonal, usually experience the highest rates of unemployment. Durable-goods manufacturing is highly cyclical, so its unemployment rate jumps in an economic downturn. From the fourth quarter of 2000 to that of 2002, total payroll employment declined 1.4 million, but in manufacturing the drop was 1.9 million, with 1.5 million in durable goods. (By contrast, employment in service-producing industries rose 0.6 million in the same period.)

Government workers have the greatest job security. Even they, however, face rising unemployment. Their unemployment rate jumped from 2.4 percent in May 2003 to 3.5 percent in June, as state and local governments, suffering from reduced tax revenues, have cut services. Although unemployment rates for workers in financial, educational, and health services have been rising, overall they are less vulnerable to unemployment, since the demand for these types of services tends to be steady.

There are very pronounced differences in susceptibility to unemployment by occupation, some of which is due to the state of the industries in which those occupations are employed and some to the supply and demand for certain skills. In June, when the national unemployment rate was 6.4 percent, the seasonally unadjusted rate for production, transportation, and materials-moving occupations was 8.7 percent, because many people in those occupations are employed in the manufacturing sector, which is hurting. During the same month, by contrast, the unemployment rate for managers and professionals was 3.4 and 3.6 percent, respectively.

Occupational differentials, in turn, are linked to variations by level of education. In June unemployment rates, seasonally adjusted, were 9.7 percent for those with less than a high school diploma, 5.8 percent for high school graduates with no college, 4.9 percent for those with some college or an associate degree, and 3.1 percent for those with a bachelor’s degree or higher.

Much of the difference in unemployment rates in different geographic locations is related to the type of economic activities that predominate in each area, but some may also result from local events. Many manufacturing centers, especially northwestern states such as Oregon, which posted an 8.3 unemployment rate in June, have seen jobs evaporate at record rates, while the current high-tech meltdown has been particularly devastating to California’s Silicon Valley. New York City, where the unemployment rate was 8.1 percent in June, has been hit hard by the Wall Street decline and also continues to suffer the effects of September 11.

Unemployment stems from three conditions affecting the labor market: the time it takes to find a new job, structural changes in the economy, and the business cycle. Based on these causes, unemployment is categorized as either frictional, structural, or cyclical.

Even in boom times some people are unemployed. In a dynamic economy, consumers and businesses buy more of some things today than yesterday and less of others, productivity changes at different rates among industries, and workers enter and leave the labor market or quit their present job to seek a better one. If there were perfect knowledge and perfect mobility in the labor market, a worker leaving one job would immediately report to a new one, but in reality it takes time to find a job, even when one exists. Frictional unemployment is unemployment that exists because labor markets are dynamic and information flows are imperfect, so it takes time for job seekers to connect with a vacant job.

A certain amount of unemployment is recurrent and periodic within one year and is considered normal. Seasonal unemployment (at one time studied as a separate category but now categorized under frictional unemployment) stems from the buying habits of the public, which are related to the weather. Most people vacation in the summer and children’s camps and many seaside resorts are open only then, while ski resorts operate only in the winter. In the colder climates, outdoor construction is less common in the winter, and the sowing-harvest cycle determines seasonal patterns in agriculture.

Some unemployment arises from structural changes in the economy causing a mismatch between the labor supply and employers’ demands, even when overall unemployment is low. In the good times following World War II, for instance, miners in northeastern Pennsylvania who were displaced from their jobs as the nation shifted from anthracite coal to oil for home heating became structurally unemployed, since no alternative jobs were available for them in that region.

Structural unemployment is often prompted by technical advances leading to new labor-saving processes, new ways of organizing enterprises, or obsolescence of products as new ones gain dominance. New skills are needed that some workers lack, and those displaced from existing jobs because of shifts in consumer tastes, imports, or company mergers may not reside near other businesses that need workers. Society can take actions to counter structural unemployment, e.g., by providing retraining and relocation programs.

Unemployment associated with the business cycle—the expansion and contraction of economic activity that occurs over the years—is termed cyclical. The degree to which an industry is exposed to cyclical fluctuations depends largely on the income elasticity of demand for its products. The easier it is for customers to postpone purchases of those products, the more income-elastic is the demand. The investment-goods industries are the most seriously affected by recession, because as sales fall, facilities operate at below capacity and companies are less inclined to invest in new plant and equipment. Such investment declined 5.2 percent in 2001 and a further 5.7 percent in 2002. The durable-goods manufacturing industries, almost by definition, are also sensitive to business cycles in that the purchase of their products can be postponed.

Society used to assume that depressions had to run their course. Now, largely because of the “Keynesian revolution,” we recognize that economic depressions are due basically to a deficiency in overall demand for products and services and can be remediated through fiscal and monetary policy. As a result, while we have not eliminated the business cycle, we have succeeded in mitigating the length and depth of economic downturns.

The American economy has undergone profound change in recent decades, with inevitable repercussions in the labor market. In the 1970s, economic growth slowed, inflation soared to the double-digit level, and the rate of productivity improvement plummeted. In the late 1970s, we began to deregulate, introducing competition into industries such as airlines and trucking. Monetary policy brought roaring inflation to a halt, but at the price, in 1981-1983, of the worst recession since World War II. Recovery, however, did not lead to a return to former conditions. Stable markets had disappeared, and companies now faced intensified domestic and foreign competition. New computer-based technology upset traditional organizational structures and the nature of jobs. Nevertheless, the nation adjusted and prosperity returned. Following a recession early in the 1990s, the rest of the decade saw rapid growth. But the structural changes in the economy led to changes in the labor market that had considerable impact on employment.

By the 1980s, certain new trends in the labor market had already become discernable. More advanced technology was shifting demand toward workers with higher-level skills, but at the same time, a greater flow of immigration was raising the supply of unskilled labor and restricting job opportunities for less skilled native workers. The labor market was also becoming more international as the economy began to globalize.5

On the bright side, in the 1990s we achieved what had been considered impossible—a combination of low unemployment and low inflation. Unlike many European countries, we continued to generate jobs, enough of them to accommodate a growing labor force even as the rate of productivity improvement rose. From 1991 to 2000, the number employed increased 19 million and the demand for unskilled labor rose, but it was high-skilled occupations—managerial, professional, and technical—that grew the fastest. Overall, the labor market adjusted successfully to the vast structural changes, albeit with difficulties for many individuals.

As a result of intensified competition and the inability to raise prices, the business enterprise has come under tremendous pressure to reduce costs. Many companies have responded by cutting jobs. These downsizings have not always been a reaction to business declines, but rather a means of improving operating efficiency. Following the heavy layoffs during the 1981-1983 recession, for example, downsizings continued even as the economy picked up, a reflection of the continuing struggle to remain competitive. From 1993 to 1995, when the economy was growing rapidly, employees were let go at almost the same rate as in the 1981-1983 deep recession. Unlike in the past, these downsizings included managerial and other white-collar employees.

Contrary to some charges, companies have not used downsizing to replace mid-career employees with younger, cheaper ones. The truth is that when reducing the number of jobs, companies have usually adhered to a traditional “LIFO” (last in, first out) seniority policy. Combined with generous severance pay and retirement inducements offered older employees, that policy has actually provided middle-career employees with more protection than others have received. Greater use of temporary help has enabled companies to reduce the extent of downsizing. The changes in employment policy, however, have weakened the bonds between workers and firms.

A number of factors explain management’s shift in attitude toward employment. Foremost, as we have shown, has been global and domestic competition, which has restricted companies’ freedom to set prices. Peter Capelli believes that democratization of the corporation has also played a role. For most of the twentieth century, the large corporation was controlled by management, not owners, and policy was fashioned to promote the long-run growth of the organization, not necessarily to enhance short-term gains for stockholders. Then, boards of directors began to use stock options based on results as a tool for motivating executives, and management began to act as shareholders. The stock market, moreover, reacts positively to corporate attempts to operate more efficiently, and an announcement of downsizing often leads to a jump in the price of a company’s stock. Another form of democratization, the flattening of the organization structure by pushing decision making down the line, has permitted the elimination of layers of management

Despite these significant shifts in the labor market, the data reveal a surprising degree of stability. It is true that job security has declined, particularly for white-collar workers, and that job losers tend to have a harder time finding a new position than in the past. But job stability, that is, the duration of jobs, is a mixed picture. The share of workers with less than ten years’ tenure has increased, but only modestly. Average job tenure, on the other hand, has actually increased slightly over the past 20 years. According to BLS, median years of employment for those employed in January 1983 was 3.5, and in January 2002, 3.7. There are demographic differences, with men and black workers experiencing less stability than women and whites. As economic growth continued in the latter part of the 1990s, job losses declined and more displaced workers were able to find new comparable employment.

Alternative employment arrangements have become an important feature of the labor market. The contingent labor force, made up of temporary employees, part-timers, and contract workers, provides a source of labor to meet short-term needs and helps preserve the jobs of regular workers in downturns, since they can be let go first. Most temporaries and part-timers work these arrangements by choice. In fact, according to BLS, only 20 percent of the 24 million part-time workers accept part-time jobs because they cannot find full-time work.

Answering the Riddles

Now that we have a better understanding of the forces affecting the labor market and how employment and unemployment are defined and measured, we are in a position to begin answering the riddles posed when we started. First, where have all the jobless gone? How do we explain that tens of thousands of jobs have disappeared without a concomitant rise in unemployment?

The answer turns on the definition of unemployment. Many of those whose jobs evaporated are not counted as unemployed, either because they have actually decided to stop seeking work for one reason or another and so have dropped out of the labor force, or because the BLS definition of unemployed excludes them even if they still see themselves as wanting to work. Some have taken part-time or temporary jobs to tide them over, and so are deemed to be employed.

To answer riddle #2, how can an economic recovery not produce new jobs, we need to factor the effects of productivity improvements into what we have learned about labor economics. Normally, as the economy begins to grow after a recession, employment picks up and, after a time lag of some months, unemployment starts to drop. But since the last recession, which ended in November 2001, jobs have continued to disappear.

The problem is that the economy is not expanding at a pace that will accommodate the increase in productivity and the growth of the labor force. In 2002, labor productivity in the nonfarm business sector increased 5.4 percent, while GDP went up only 2.4 percent. Enhanced productivity, that is, achieving more output per unit of input, is the source of a nation’s rising living standard, and therefore highly desirable over the long run, but it also reduces the need for workers in the short run. Contributing to the problem is that more manufactures are coming from abroad.

Suffering from a deficiency in demand, which is estimated to be equal to about 2 percent of GDP, the economy is operating below its potential. Although the worst seems to be over, below-par growth may persist. The world’s other major economies, those of Japan and the European Union, are performing worse than our own, and as a result are buying fewer of our products. (Even they have been showing some signs of life, however, which augers for a better future.) The Federal Reserve has acted aggressively to remedy the situation, slashing interest rates drastically. President Bush’s “supply-side” strategy focusing on long-term improvements in economic efficiency has not been successful so far in stimulating rapid growth. His newest tax cut, now going into effect, is also supply-side oriented, but it contains demand-boosting provisions. We shall have to wait and see if they lead to an economy operating close to its potential.

Even getting economic growth back to its trend rate will not immediately eliminate excess unemployment. As jobs again become available, some of the people who had postponed entering the labor market will do so, and others who dropped out of the labor market will come back. If the rate of labor force participation were to return to prerecession levels, 1.2 million more job seekers would be competing for work. We have to conclude, then, that not only is the outlook for rapid economic growth problematic, but getting back to a 4 percent unemployment rate is only a distant possibility.

This phenomenon is not new. Two decades ago, Summers found that almost half of all unemployment spells end by persons leaving the labor force.
Further complicating matters is the fact that in May of this year BLS implemented several changes to its establishment survey, including a new industrial classification system designed to better reflect today’s economy (e.g., by breaking out an information sector), concurrent seasonal adjustment, and completion of the conversion to a probability-based sample.
How long people are unemployed depends on the state of the labor market. In June of this year, in a loose labor market, only 31.6 percent of the unemployed were unemployed for less than five weeks. Another 30.9 percent were unemployed for five to 14 weeks, and 37.5 percent for 15 weeks or longer. Two million, or 21.4 percent of all the unemployed, had been jobless for more than half a year. Even those figures do not provide a full picture of unemployment’s impact on individual workers, because some suffer from more than one spell during a year.
Separate unemployment data for Asians have been available only since 2000.
For an illustration of the impact of globalization on labor economics, see “Sweatshops in the Garment Industry: Lessons in Global Development,” IRConcepts, Summer 2003.

  • Berle, A.A., and Gardiner Means, The Modern Corporation and Private Property, Macmillan, 1932.
  • Capelli, Peter, “Examining the Incidence of Downsizing and its Effect on Establishment Performance,” in Neumark, op.cit.
  • Helfgott, Roy B., “Labor Market Models in Europe and America—and Unhappiness with Both,” Business Horizons, March–April 1996.
  • Neumark, David, ed., On the Job: Is Long-term Employment a Thing of the Past? Russell Sage Foundation, 2000.
  • Summers, Lawrence H., Understanding Unemployment, MIT Press, 1990.
    Various reports and articles in the Monthly Labor Review, U.S. Department of Labor, Bureau of Labor Statistics.