Macroeconomics Chapter 6: Unemployment Sixth Edition N. Gregory Mankiw ECON 4020/Chatterjee

In this chapter, you will learn… …about the natural rate of unemployment: what it means what causes it understanding its behavior in the real world CHAPTER 6 Unemployment

Natural rate of unemployment The unemployment rate: the fraction of the civilian workforce that is unemployed The “Natural” rate of unemployment: The average or long-run rate of unemployment around which the economy fluctuates. In a recession, the actual unemployment rate rises above the natural rate. In a boom, the actual unemployment rate falls below the natural rate. The natural rate of unemployment is the “normal” unemployment rate the economy experiences when it is neither in a recession nor a boom. CHAPTER 6 Unemployment

Actual and natural rates of unemployment in the U.S., 1960-2006 12 Unemployment rate 10 8 Percent of labor force Natural rate of unemployment 6 4 Figure 6-1, p.160. The actual unemployment rate fluctuates considerably over the short run. These fluctuations will be the focus of chapters 9-13 later in the book. For this chapter, though, our goal is to understand the red line: the so-called “natural rate of unemployment,” or the long-run trend in the unemployment rate. Source: BLS Obtained from http://research.stlouisfed.org/fred2/ Unemployment data are based on seasonally-adjusted, monthly unemployment rates for the civilian non-institutional population of the U.S. The actual u-rate for each quarter is an average of the three monthly unemployment rates in that quarter. The natural u-rate in a given quarter is estimated by averaging all unemployment rates from 10 years earlier to 10 years later; future unemployment rates are set at 5.5%. 2 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 CHAPTER 6 Unemployment

A first model of the natural rate Notation: L = workers in labor force E = employed workers U = unemployed workers U/L = unemployment rate Section 6-1 CHAPTER 6 Unemployment

Assumptions: 1. L is exogenously fixed (no population growth). 2. During any given month, s = fraction of employed workers that become separated from their jobs s is called the rate of job separation f = fraction of unemployed workers that find jobs f is called the rate of job finding s and f are exogenous CHAPTER 6 Unemployment

The transitions between employment and unemployment s E Employed Unemployed f U Figure 6-2, p. 161 (note: The size of the boxes containing the words “employed” and “unemployed” are not proportional to the number of people in each category.) CHAPTER 6 Unemployment

The steady state condition Definition: the labor market is in steady state, or long-run equilibrium, if the unemployment rate is constant. The steady-state condition is: s E = f U # of unemployed people who find jobs # of employed people who lose or leave their jobs CHAPTER 6 Unemployment

Finding the “equilibrium” unemployment rate f U = s  E = s  (L – U ) = s L – s U Solve for U/L: (f + s) U = s L so, CHAPTER 6 Unemployment

The “Natural” Rate of Unemployment Therefore, the natural rate of unemployment is defined as: CHAPTER 6 Unemployment

Example: Each month, Find the natural rate of unemployment: 1% of employed workers lose their jobs (s = 0.01) 19% of unemployed workers find jobs (f = 0.19) Find the natural rate of unemployment: CHAPTER 6 Unemployment

Policy implication A policy will reduce the natural rate of unemployment only if it lowers s or increases f. CHAPTER 6 Unemployment

Why is there unemployment? If job finding were instantaneous (s = 0, f = 1), then all spells of unemployment would be brief, and the natural rate would be near zero. There are two reasons why f < 1 and s > 0: 1. job search 2. wage rigidity CHAPTER 6 Unemployment

Job search & frictional unemployment frictional unemployment: caused by the time it takes workers to search for a job occurs even when wages are flexible and there are enough jobs to go around occurs because workers have different abilities, preferences jobs have different skill requirements geographic mobility of workers not instantaneous flow of information about vacancies and job candidates is imperfect CHAPTER 6 Unemployment

Sectoral shifts def: Changes in the composition of demand among industries or regions. example: Technological change more jobs repairing computers, fewer jobs repairing typewriters example: A new international trade agreement labor demand increases in export sectors, decreases in import-competing sectors Result: frictional unemployment Sometimes the unemployment caused by sectoral shifts is severe. Due to increasing imports of cheaper foreign-made textiles (particularly since the expiration in 2005 of long-standing quotas on textiles from China), the U.S. textile industry has been in decline for years. Tens of thousands of workers in this industry have lost jobs. Many of these workers are in their 50s and have worked in this industry for decades. Such workers are unlikely to have the skills necessary to get jobs available in newly booming industries, and they are less likely to invest in the acquisition of the necessary skills for these jobs. Hence, such workers are at greater risk for becoming “discouraged workers.” CHAPTER 6 Unemployment

CASE STUDY: Structural change over the long run All figures are industry shares in U.S. GDP. “Other industry” includes construction, mining, electricity, water, and gas. From 1960 to 2000, there are huge changes in all four categories. Manufacturing falls by about a third. Even the “tiny” category of agriculture drops by nearly two-thirds: from 4.2% to 1.6% of GDP. These changes represent HUGE structural shifts, which vastly alter the kinds of jobs in demand. Source: World Development Indicators, World Bank. CHAPTER 6 Unemployment

More examples of sectoral shifts Late 1800s: decline of agriculture, increase in manufacturing Late 1900s: relative decline of manufacturing, increase in service sector 1970s: energy crisis caused a shift in demand away from gas guzzlers toward smaller cars. Most of the examples on this and the previous slides are big changes that have occurred over many years. These examples give students a good idea of what sectoral shifts are. Perhaps more important for the natural rate, though, are the many smaller changes that occur more frequently. Ours is a dynamic economy: the structure of demand is shifting almost continuously, due to changes in preferences, technology, and the location of production. As a result, there is a near-continual flow of newly frictionally unemployed workers. Sectoral shifts are distinct from recessions (which also cause unemployment). In recessions, there is a general fall in demand across industries, and the unemployment that results is cyclical. Sectoral shifts, though, are changes in the composition of demand across industries, and lead to frictional unemployment as described above. CHAPTER 6 Unemployment

Sectoral shifts U.S. Auto Industry (Detroit): Struggling for many years: falling profits, increased foreign competition, insufficient R&D - Recently, rising healthcare costs have caused significant increase in costs -Auto workers have been leaving the industry to work in the services sector…healthcare (especially nursing) is an attractive option -But transition can be painful (new skills, training, going back to school, etc) In a dynamic economy, sectoral shifts occur frequently, contributing to frictional unemployment: often referred to as the “cost” of economic development CHAPTER 6 Unemployment

Geography of a Recession U.S. Housing and Labor Markets January 2008 CHAPTER 6 Unemployment

Public policy and job search Govt programs affecting unemployment Govt employment agencies: disseminate info about job openings to better match workers & jobs. Public job training programs: help workers displaced from declining industries get skills needed for jobs in growing industries. You might want to “hide” (omit) this slide from your presentation if you plan on doing the class discussion in Slide 27, which asks students to think of things the government can do to try to reduce the natural rate of unemployment. CHAPTER 6 Unemployment

Unemployment insurance (UI) UI pays part of a worker’s former wages for a limited time after losing his/her job. UI increases search unemployment, because it reduces the opportunity cost of being unemployed the urgency of finding work Rate of job finding, f Studies: The longer a worker is eligible for UI, the longer the duration of the average spell of unemployment. The text includes a nice case study on unemployment insurance (pp.164-165). It discusses evidence that unemployment insurance reduces the job finding rate. CHAPTER 6 Unemployment

Benefits of UI By allowing workers more time to search, UI may lead to better matches between jobs and workers, which would lead to greater productivity and higher incomes. CHAPTER 6 Unemployment

Why is there unemployment? The natural rate of unemployment: Two reasons why f < 1: 1. job search 2. wage rigidity DONE  Next  CHAPTER 6 Unemployment

Unemployment from real wage rigidity Supply Labor Real wage If real wage is stuck above its eq’m level, then there aren’t enough jobs to go around. Unemployment Demand Rigid real wage Amount of labor hired Figure 6-3 on p.166. Abbreviation: “eq’m” = equilibrium Amount of labor willing to work CHAPTER 6 Unemployment

Unemployment from real wage rigidity If real wage is stuck above its eq’m level, then there aren’t enough jobs to go around. Then, firms must ration the scarce jobs among workers. Structural unemployment: The unemployment resulting from real wage rigidity and job rationing. Other texts define “structural unemployment” as unemployment that results from a mismatch between the skills or locations of workers and the skill or location requirements of job openings. This would occur, for example, if there were a decrease in demand for domestic steel (and hence steel workers) and a simultaneous increase in demand for financial consulting services (and hence employees of such firms). However, if wages are perfectly flexible, then the decrease in demand for steel workers would simply cause their wage to fall until all were again employed, and the increase in demand for workers in financial firms would simply increase until equilibrium in that labor market was reestablished. So, the critical ingredient for structural unemployment is wage rigidity. Hence Mankiw’s definition. CHAPTER 6 Unemployment

Reasons for wage rigidity 1. Minimum wage laws 2. Labor unions 3. Efficiency wages CHAPTER 6 Unemployment

1. The minimum wage The min. wage may exceed the equilibrium wage of unskilled workers, especially teenagers. Studies: a 10% increase in min. wage reduces teen employment by 1-3% Tendency for firms to substitute towards illegal workers (who are not bound by the min. wage) But, the min. wage cannot explain the majority of the natural rate of unemployment, as most workers’ wages are well above the min. wage. CHAPTER 6 Unemployment

2. Labor unions Unions exercise monopoly power to secure higher wages for their members (collective bargaining). When the union wage exceeds the equilibrium wage, unemployment results. Insiders: Employed union workers whose interest is to keep wages high. Outsiders: Unemployed non-union workers who would be willing to work for lower wages, so there would be enough jobs for them. See p.165 for more discussion about insiders and outsiders. The theory has two implications we can confront with data: 1) Union members’ average earnings should be higher than non-union members’ average earnings. 2) The difference between union and non-union wages should be higher in industries that are more heavily unionized (and hence, in which unions have more market power) than in less heavily unionized industries. The following slide shows 2005 data on union membership and wage ratios by industry in the U.S. The data are consistent with the theory. CHAPTER 6 Unemployment

Union membership and wage ratios by industry, 2005 # employed (1000s) U % of total wage ratio Private sector (total) 105,508 8.5% 40.5 8 15.4 3.1 2.1 24.4 5.8 13.7 9.5 13.8 122.3 121.7 115.1 112.7 90.6 90.7 129.2 114.0 107.8 113.7 156.9 Government (total) 20,381 Construction 8,053 Mining 600 Manufacturing 15,518 Retail trade 14,973 Transportation 4,379 For this slide, “union members” includes workers that are in a union or similar worker association, or whose jobs are covered by a union contract. U % of total = “union members” (as defined above) as a percentage of all workers Wage ratio = average weekly earnings of “union members” (as defined above) as a percentage of average weekly earnings of nonunion workers. For example, in the transportation industry, 24.4% of workers are in unions and earn 29.2% more than non-union workers in this industry. Source: BLS.gov Note: Due to space constraints on the slide, some industries were omitted. In 2005, about 13% of all workers in the U.S. were members of unions. The data on this slide show two things: 1) union workers typically earn more than non-union workers (about 22-23% more on average). 2) the greater the percentage of union workers in an industry, the higher the wage ratio is likely to be (the correlation is about 0.5) Finance, insurance 6,304 Professional services 10,951 Education 3,312 Health care 14,045 wage ratio = 100(union wage)/(nonunion wage) slide 28

3. Efficiency wage theory Idea: higher wages increase worker productivity by: Attracting higher quality job applicants (“Adverse Selection” problem) Increasing worker effort, reducing “shirking” (“Moral Hazard” problem) Reducing turnover, which is costly to firms Improving health of workers: better nutrition & productivity (in developing countries) Firms willingly pay above-equilibrium wages to raise productivity, causing structural unemployment. CHAPTER 6 Unemployment

Question for discussion: Use the material we’ve just covered to come up with a policy or policies to try to reduce the natural rate of unemployment. Note whether your policy targets frictional or structural unemployment. It is useful to pause your lecture at this point and give students an opportunity to apply what you’ve covered so far to answer this policy question. Possible answers: Stop raising the (nominal) minimum wage, so that its real value will gradually erode to zero. Regulate unions (just like other monopolies are regulated) to reduce unions’ impact on wages. Reduce the generosity of unemployment insurance benefits. Implement government employment agencies to increase the accessibility of information about job vacancies and available workers. Increase public funding to help retrain workers displaced from jobs in declining industries. Suggestions for conducting the discussion: If you ask for responses immediately after posing the question, it is likely that a small number of students will volunteer to participate - the same students that always do, the ones that are the best prepared and/or the quickest thinkers. To elicit participation from a larger number of students, I suggest the following: Pair students up. Allow 10 minutes for the students, working in their pairs, to come up with answers to the question. During this time, circulate around the room and ask the pairs if you can be of assistance, either to help them get started or give feedback on what they’re coming up with. Then, reconvene the class and ask for volunteers. Doing this increases the quantity and quality of participation: students who would not otherwise participate are more likely to do so because they have had time to formulate their answers and have had a chance to run their answers by a classmate. Additionally, even students who don’t participate will have at least had the opportunity to discuss the question with one other student. CHAPTER 6 Unemployment

The duration of U.S. unemployment, average over 1/1990-5/2006 # of weeks unemployed # of unemployed persons as % of total # of unemployed amount of time these workers spent unemployed as % of total time all workers spent unemployed 1-4 38% 7.2% 5-14 31% 22.3% 15 or more 70.5% Source: Bureau of Labor Statistics (www.bls.gov) and author’s calculations. How to interpret this data: The second column shows the percentage of all unemployed workers whose spell lasted the number of weeks shown (average over January 1990 to May 2006). The third column shows the share of total time spent unemployed attributed to workers in that category. I calculated it as a ratio. The denominator is the total number of weeks spent unemployed, obtained by multiplying the total number of unemployed persons by the number of weeks of the average spell of unemployment. The numerator is, for each category, the number of people in that category times the duration of the average spell of unemployment for that category. (I assume the duration of the average spell is the midpoint of each category, and 30 weeks for the “15 or more” category. I have tried different assumptions and the results are similar to those shown.) 7.2% of total time spent unemployed was spent by people who were unemployed for less than 5 weeks. 70.5% of total time spent unemployed is attributed to people who were unemployed for 15 or weeks or longer. The point of this data: More spells of unemployment are short term (38%) than medium term (31%) or long term (31%). But, most of the time spent unemployed is long-term. CHAPTER 6 Unemployment

The duration of unemployment The data: More spells of unemployment are short-term than medium-term or long-term. Yet, most of the total time spent unemployed is attributable to the long-term unemployed. This long-term unemployment is probably structural and/or due to sectoral shifts among vastly different industries. Knowing this is important because it can help us craft policies that are more likely to work. Regarding the point about structural unemployment and sectoral shifts: Structural unemployment - workers are waiting for jobs to become available, but there just aren’t enough jobs to go around; hence, this can be long-term unemployment. Sectoral shifts across vastly different industries, e.g. a shift in demand from textiles to software design; obviously, jobs in these industries require vastly different skill sets. Sectoral shifts can occur among similar industries (e.g., demand shifts from desktop to laptop computers), but this is less likely to produce long-term unemployment. CHAPTER 6 Unemployment

TREND: The natural rate rises during 1960-1984, then falls during 1985-2006 The purpose of this slide is to establish the trend behavior of the natural rate in recent decades: rising until the early 80s, then falling from the mid-80s through the early 2000s. The following slides will show that the theories in this chapter are (mostly) consistent with the trend behavior of the natural rate. The graph on this slide is similar to Figure 6-1 (near the beginning of this PowerPoint presentation). However, since we are now focusing on the trend behavior in the natural rate, I have altered the vertical axis to make the trend more apparent, and I’ve dimmed the actual unemployment rate data – now light gray. CHAPTER 6 Unemployment

EXPLAINING THE TREND: The minimum wage The trend in the real minimum wage is similar to that of the natural rate of unemployment. 9 8 7 6 minimum wage in 2006 dollars 5 Dollars per hour 4 The trend in the real minimum wage rises until the mid to late 1970s, then falls. This is fairly similar to the trend of the natural rate of unemployment. The U.S. Department of Labor has lots of good information on the minimum wage, at: http://www.dol.gov/dol/topic/wages/minimumwage.htm 3 minimum wage in current dollars 2 1 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 CHAPTER 6 Unemployment

EXPLAINING THE TREND: Union membership Union membership selected years year percent of labor force 1930 12% 1945 35% 1954 1970 27% 1983 20.1% 2005 12.5% Since the early 1980s, the natural rate of unemploy-ment and union membership have both fallen. But, from 1950s to about 1980, the natural rate rose while union membership fell. source: AFL-CIO website http://www.aflcio.org and, for the 2005 figure, www.BLS.gov Also see http://www.lib.umich.edu/govdocs/steclab.html#unions good set of links to info on labor unions http://stats.bls.gov/news.release/union2.toc.htm The BLS’ annual news release of info on union membership, earnings. Earlier in the chapter, we saw cross-sectional data that showed a positive correlation between the union wage premium and union members’ share of the labor force across industries. We would expect that, other things equal, changes over time in the aggregate share of unions in employment should be associated with similar changes in unions’ impact on average wages, and hence on the natural rate of unemployment. In plain English, we would expect that the decline in the extent of unionization shown on this slide would correspond to a decline in the natural rate of unemployment. Unfortunately for the theory, this is only true for the time period beginning in the early 1980s, when the natural rate started coming down. From the 1950s through 1980, the natural rate rose, but union membership fell. Does this mean the theory is not relevant? Not necessarily, as other things (other determinants of the natural rate) were not constant during this time period. CHAPTER 6 Unemployment

EXPLAINING THE TREND: Sectoral shifts From mid 1980s to early 2000s, oil prices less volatile, so fewer sectoral shifts. Price per barrel of oil, in 2006 dollars source: Dow Jones & Company obtained from: http://research.stlouisfed.org/fred2/ Earlier in the chapter, we learned that sectoral shifts are a source of job separations and lead to frictional unemployment. One would expect that a decrease in the frequency and magnitude of sectoral shifts would be associated with fewer job separations, less frictional unemployment, and a lower natural rate of unemployment. Unfortunately, there is no single “index of sectoral shocks.” However, we know that large changes in oil prices are one source of sectoral shocks. A significant fall in the price of oil causes a decrease in demand for workers at oil fields in Oklahoma and Texas, and an increase in demand for workers at factories that produce SUVs. A significant increase in oil prices would do the opposite. The graph shows data on the price of oil since 1970. During 1970-1985, the real price of oil fluctuated between $18 and $98. Also during this time, the natural rate of unemployment was rising. During 1986-2002, the real price of oil fluctuated between $20 and $40, except for a brief spike during the Gulf War. Also during this time, the natural rate of unemployment was falling. The data are roughly consistent with the notion that sectoral shifts contribute to the natural rate. Note the recent increase in oil prices: from about $22 to $70 during 2002-2006:1. This represents a sectoral shift and may contribute to an increase in the natural rate of unemployment. Or maybe not, as oil consumption per dollar of GDP is lower today than in the 1970s and 1980s. CHAPTER 6 Unemployment

EXPLAINING THE TREND: Demographics 1970s: The Baby Boomers were young. Young workers change jobs more frequently (high value of s). Late 1980s through today: Baby Boomers aged. Middle-aged workers change jobs less often (low s). For details on this, plus one other explanation involving productivity, see pp.174-175. CHAPTER 6 Unemployment

Unemployment in Europe, 1960-2005 France 12 9 Percent of labor force 6 Italy U.K. Figure 6-4, p.177 Source: bls.gov, obtained from http://www.bls.gov/fls/home.htm 3 Germany 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 slide 38

The rise in European unemployment Shock Technological progress has shifted labor demand from unskilled to skilled workers in recent decades. An increase in the “skill premium” – the wage gap between skilled and unskilled workers. Higher unemployment, due to generous govt benefits for unemployed workers strong union presence. CHAPTER 6 Unemployment

Percent of workers covered by collective bargaining United States 18% United Kingdom 47 Switzerland 53 Spain 68 Sweden 83 Germany 90 France 92 Austria 98 Table 6-1, p.169 Source: Same as text (OECD Economic Outlook 2004, as reported in NBER Macroeconomics Annual 2005.) CHAPTER 6 Unemployment

Chapter Summary 1. The natural rate of unemployment the long-run average or “steady state” rate of unemployment depends on the rates of job separation and job finding 2. Frictional unemployment due to the time it takes to match workers with jobs may be increased by unemployment insurance CHAPTER 6 Unemployment slide 41

Chapter Summary 3. Structural unemployment results from wage rigidity: the real wage remains above the equilibrium level caused by: minimum wage, unions, efficiency wages 4. Behavior of the natural rate in the U.S. rose from 1960 to early 1980s, then fell possible explanations: trends in real minimum wage, union membership, prevalence of sectoral shifts, and aging of the Baby Boomers CHAPTER 6 Unemployment slide 42

Chapter Summary has risen sharply since 1970 5. European unemployment has risen sharply since 1970 probably due to generous unemployment benefits, strong union presence, and a technology-driven shift in demand away from unskilled workers CHAPTER 6 Unemployment slide 43