Chapter 3 Understanding unemployment 0 Why you should be interested in unemployment Unemployment is closely related to the emergence of capitalism But mass unemployment as a permanent phenomenon only appeared in the 1970 s following the oil shocks Of course, even mass unemployment only affects a small part of the workforce Unemployment rate goes from 2% to 12% in OECD countries (7% in Canada) Unemployment tends to be concentrated on risk groups (low-skilled, entrants, women ) 1
Why you should be interested in unemployment Unemployment is challenging at the individual level He/she is stigmatized Income falls Credit is not available Health risks are higher But also at the family level Children stay in the family longer Unemployed lower family income 2 Why you should be interested in unemployment And at the economywide level Unemployment means poverty and inequality Youth criminality is higher Unemployment compensation/welfare costs increase (tax rates must increase) 3
Why you should be interested in unemployment Finally, the youth are the main victims of unemployment Don t worry: education protects from unemployment! 4 Why you should be interested in unemployment Male ratio of unemployment rate by age group to average unemployment for 8 OECD counrtries, 1997 5
Understanding unemployment Given the persistance of unemployment rates at very high levels, this chapter is devoted to structural unemployment 6 Outline of the chapter 1. The natural rate of unemployment 2. Job search and market frictions 3. Rigid real wages 4. Case study: the European unemployment problem 7
1. The natural rate of unemployment 8 Definitions The natural rate of unemployment is (Theoretical) the rate of unemployment that still exists in the economy when the labour market is in equilibrium (Statistical) the average rate around which the economy fluctuates During a recession, the actual unemployment rate goes above the natural rate During booms, the actual unemployment rate falls below the natural rate
Computing the natural rate There are three types of methods Natural rate = mean value of unemployment over a given period of time Natural rate = trend component of unemployment U t = C t + T t Natural rate = Non-Accelerating Inflation Rate of Unemployment 10 Natural rate and NAIRU Consider the following simple Phillips curve π t - π t-1 = α - βu t + ε t where π is the rate of inflation, α and β are parameters and ε is an error term with an expected value of 0 π t-1 is the assumed expected rate of inflation. There is no long run monetary illusion 11
Natural rate and NAIRU The NAIRU is the rate compatible on average with constant inflation rate π t = π t-1 implies U = α / β Richer estimates of the NAIRU incorporate longer lags on the inflation and unemployment rates, control for episodes of supply shocks and wage/price controls Typically α depends on time 12 Time-varying unemployment rate The natural unemployment rate results from structural factors such as wage bargaining rules, the generosity of unemployment insurance, tax rates, minimum wage As those factors change over time, the natural unemployment rate is not necessarily constant Time-varying natural unemployment rate 13
US unemployment rate, 1958-2002 11 10 9 8 7 6 5 4 3 2 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 To compute the natural rate, you need to extract the trend component of actual unemployment rate 14 Flows, stocks and the natural rate Notation: N = Workforce E = Employment U = Unemployment U/N = Unemployment rate Of course, E +U = N 15
Assumptions 1. N is fixed: the aggregate labour supply is perfectly inelastic, and there is no entry on the labour market 2. Each month, q = proportion of employees losing their job, p = proportion of unemployed finding a job Parameters p and q determine the speed of transitions between the different states of the labour market 16 To fix the ideas Country US France Japan UK Germany Entry rate q 2.06 0.34 0.38 0.67 0.57 Exit rate p 37.4 3.4 17.1 9.3 9.0 Unemployment: entry and exit rates, 1993 (in %) 17
Transitions between employment and unemployment q E Employment Unemployment p U 18 Long run equilibrium Finding the natural rate is finding stationnary populations in each state, ie within Employment and Unemployment boxes Natural rate = long run equilibrium or stationary equilibrium 19
Long run equilibrium Definition: the long run equilibrium of the labour market is characterized by the equality of flows in and out of unemployment Formally: q E = p U Number of workers who quit or lose their job Number of unemployed who find a job 20 In quest of the long run equilibrium p U = q E = q (N U ) = q N q U Isolating terms in U and N: (p + q) U = q N Hence, q Natural unemployment rate = U = N q + p 21
To fix the ideas Countries Entry rate q Exit rate p Natural rate US 2.06% 37.4% 5.2% France 0.34 3.4 9.1 Japan 0.38 17.1 2.2 UK 0.67 9.3 6.7 Germany 0.57 9.0 6.0 U q = N q + p 22 Things to notice Very different labour markets may have the same unemployment rate It is possible to sustain a given unemployment rate with either high transition rates, or low transition rates High p/high q vs low p/low q Example: Australia vs Canada 23
Where do you want to live? 1990 1995 1999 Australia 7/41 8.5/54.1 7.2/48.4 Canada 8.1/20.2 9.5/29.1 7.6/21.4 Unemployment rate and share of long-term unemployed (>6 months) among the unemployed for selected years 24 Why is there equilibrium unemployment? U q = N q + p To understand U, we must know the long run values of p and q If it was very easy to find a job (p = 1), unemployment durations would be very brief, and the natural rate would be close to 0 It was impossible to lose a job (q = 0), there would be no entry in unemployment and the natural rate would be 0 25
Why is there equilibrium unemployment? U q = N q + p The different factors behind job destruction q = q(job reallocation, individual mismatch, labour demand, employment protection) q = q JR q M q LD q EP q JR q M = regular job turnover q LD = impact of the labour demand 26 Why is there equilibrium unemployment? U q = N q + p The different factors behind job-finding rate p = p(matching efficiency, individual job search, labour demand) p = p ME p JS p LD p ME p JS = job-finding rate resulting from regular market frictions p LD = impact of the labour demand 27
Why is there equilibrium unemployment? This decomposition of entry rate and exit rate suggests two main sources of structural unemployment Labour market frictions, which explain Regular job turnover: job reallocation + individual mismatch Regular job-finding rate: matching efficiency + job search Wage rigidities, which determine the structural level of the labour demand 28 2. Labour market frictions 29
Frictional unemployment Definition: unemployment originating from regular job turnover and the regular delay separating two periods of employment 30 Losing one s job Regular job turnover consists in individual mismatch and job reallocation Individual mismatch arises whenever you (or your employer) realize your skills are not valued at best in the job you occupy Job reallocation is mainly induced by technological change 31
Job reallocation Definition: changes in the demand composition by sector or region Major changes take place over very long periods Agriculture/industry/service But, for the natural rate of unemployment, the most important changes are minor changes within industries 32 Industry shares in U.S. GDP, 1960 57.9% 9.9% Agriculture Manufacturing Other industry Services 28.0% 4.2% 33
Industry shares in U.S. GDP, 1997 72.0% Agriculture Manufacturing Other industry Services 8.5% 17.8% 1.7% 34 more examples: Sectoral shifts abound 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 huge gas guzzlers toward smaller cars. In our dynamic economy, smaller (though still significant) sectoral shifts occur frequently, contributing to frictional unemployment. 35
Job reallocation Joseph Schumpeter is very well-known for his concept of creative destruction Growth is associated to innovations. But innovations tend to substitute to earlier technologies/products. However, creative destruction needs factor reallocation, and, therefore, job loss and induced unemployment Who is the schumpeterian heros? the entrepreneur or the unemployed? 36 Finding a job Regular job-finding rate results from matching efficiency and individual job search behaviour At first glance, matching efficiency broadly denotes the speed of the process allocating unemployed workers to vacant jobs Job search behaviour is the way unemployed workers alter their chances to find a job 37
Market frictions For the society as a whole, the role of the labour market is to put The right man at the right place the labour market is a matching market However, the matching process takes time, because there are matching frictions It means unemployed workers and vacant jobs are engaged in mutual search 38 The Beveridge curve Consequently, we observe both unemployed workers and vacant jobs, even though wages are perfectly flexible The curve relating the rate of unemployment to the vacancy rate is the Beveridge curve 39
Frictional unemployment vacancy rate 1,4 1,2 1 0,8 0,6 0,4 0,2 0 0,04 0,06 0,08 0,1 0,12 0,14 0,16 unemployment rate The Beveridge curve: Unemployment rate and vacancy rate, France, 1960-1998 40 Why are there market frictions? Workers are heterogeneous skills, preferences Jobs are heterogeneous skill requirements, preferences Limited geographic mobility Fixed costs while selling/buying houses Imperfect information on available jobs and applicants 41
Matching efficiency One way to appreciate the efficiency of the matching market is to measure the vacancy rate required to sustain a given unemployment rate Another way is to estimate a matching function, relating the number of hires to the numbers of unemployed and the number of vacancies 42 Matching function A typical matching function M = M(V,U) = V α U 1-α Empirical studies on aggregate data tend to validate the constant returns to scale Cobb- Douglas technology, with a value of α between 0.5 and 0.7 43
Job search Unemployed workers differ from inactive individuals because they actively search a job Actually, unemployed workers have two important decisions to make What should be my search intensity? Number of letters, search methods When should I accept a job offer? Lowest acceptable wage 44 Search intensity Search intensity results from a standard trade-off between the marginal cost and the marginal benefit from search effort Marginal Benefit increases with expected wage does not depend on search investment Marginal Cost increases with liquidity constraints increases with search investment 45
Search intensity MC MB Search effort 46 Search intensity How can we measure search intensity? Number of letters sent to potential employers Number of search methods used Time allocated to job search 47
Reservation wage The reservation wage is the lowest acceptable wage The reservation wage depends on the waiting value of unemployment The waiting vaue of unemployment depends on Expected wage offer distribution: mean wage, variance and risk aversion Individual patience: discount rate Unemployment income: underground economy, unemployment insurance, family support Job availability: labour demand pressure 48 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 raises the value of waiting unemployment, and so tends to raise the reservation wage It reduces the urgency of finding work, and so tends to decrease search intensity
UI and search intensity MC 2 MC 1 MB Search effort 50 Main results Empirical studies - The longer a worker is eligible for UI, the longer the duration of the average spell of unemployment - The replacement ratio has ambiguous effects on search effort - UI tends to raise the reservation wage 51
3. Wage rigidities 52 Without loss of generality, the labour supply is here inelastic If the real wage is above the equilibrium wage, labour demand is lower than labour supply real wage Rigid real wage Graphically Actual employment Supply Unemployment Demand Employment Individuals willing to work 53
Why are there real wage rigidities? Real wage rigidities have two origins: institutions, and market Institutional rigidities minimum wage (studied ch 4) unions Market-induced rigidities efficiency wages 54 Union goals A union is composed of individual members. If all were identical, the union would like to maximize the economic situation of the representative member 1st goal: get the best combination of wage and job loss risk But members are heterogenous. Some are very productive, and thus expect to get high wages, while others are less productive and their expected wage is low 2nd goal: reduce wage inequality 55
Union methods How can a union get a wage higher than the market wage? Unions have some monopoly power on the labour supply in a given employment area. This monopoly power can be used to bargain wages The threat of strike Wages are then bargained between unions and employers 56 The keypoint When the bargained wage is higher than the full employment wage, unemployment rises Union workers are insiders whose interest is to obtain very high wages. They are protected against unemployment risk due to their seniority The unemployed and the non-union workers are outsiders. They would prefer a lower wage than the wage bargained by the union 57
mining Services all Union premium by industry in the USA in 2001 industry construction manufacturing Transportation Public utilities Wholesale trade Retail trade Finance, insurance Government employed ( 1000) 531 6881 18149 4441 2981 4540 20505 7648 34261 19155 119092 %unioni zed 12.3% 18.4 14.6 24.1 22.6 37.4 13.6% 12.9% 19.0 15.5 25.4 23.7 41.8 15.0% 103.4 151.0 105.9 127.8 104.2 105.8 117.8 90.1 103.3 121.1 118.0 covered = non-unionized under a collective agreement Econ Wage 360 Labour ratio Economics =100 (union wage + covered Chapter wage)/(non-union 3 wage) 58 5.5 4.5 2.1 5.9 %covere d 5.9 5.0 2.8 6.8 Wage ratio Measuring union power A first measure union power is proportional to strike involvement Union power = % of workers involved in strike French peculiarities 1968: 100% Fair decline in the 1970 s 59
Measuring union power 100,0 80,0 60,0 40,0 20,0 0,0 1950 1960 1970 1980 1990 Workers involved in strike on dependent employees, France, 1950-1998 (in %) 60 Measuring union power The bigger, the stronger : union power should rise with the number of unionized workers Union power = union density, i.e. the proportion of unionized workers 61
Measuring union power 30,0 25,0 20,0 15,0 10,0 5,0 1950 1960 1970 1980 1990 2000 Union density, France, 1950-1998 62 Measuring union power US union membership selected years Year percent of labor force 1930 12% 1945 35% 1954 35% 1970 27% 1983 20.1% 2001 13.5% Since the early 1980s, the natural rate of of unemployment and union membership have both fallen. But, from 1950s to to about 1980, the natural rate rose while union membership fell. 63
Measuring union power However, union density does not play a key role In many countries, employers are obliged to bargain wages with representative unions, whatever union density Many wages, including those of nonunionized workers, depend on a collective agreement 64 Country France UK Canada Japan US Union density 9 34 38 24 16 %workers covered by a collective agreement 95 47 36 21 18 Union membership and bargaining coverage, 1994 65
Measuring union power France Italy U-K Sweden 1980/1989 85% 85% 70% 86% 1990/1998 92% 83% 47% 89% Proportion of workers covered by a collective agreement 66 Efficiency wage theories Those theories argue employers competition can lead them to set a wage higher than full employment wage The resulting real wage rigidity is induced by the market functioning Starting point of the theories: productivity rises with wage. A higher wage Attracts more productive workers Increases effort/reduces shirking Lowers turnover Improves health (in the DCs) 67
Basic mechanism The shirking model (Shapiro and Stiglitz) Effort is imperfectly observable / verifiable Workers must be incited to make an effort How? By paying a wage higher than the other employers do Problem! All firms have the same incentives, and therefore they all pay the same wage Workers are finally incited to make the effort as unemployment rises 68 Efficiency wage theories Efficiency wage models have been criticized on the grounds there is a missing control in employers hands The wage is used as a tool to pay for labour services and to motivate workers. Without such restrictions on employers behaviour, efficiency wage theories lose their ability to explain structural unemployment 69
4. Case study: European unemployment 70 Unemployment rate 12 10 8 6 4 2 The rise in European unemployment 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 71
The French case 0,16 0,14 0,12 0,1 0,08 0,06 0,04 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 72 Which explanations? Two explanations based on labour market institutions In the 1960s, institutions were employmentfriendly. Oil shocks have turned them unemployment-friendly Institutions did not change, but the economic environment did European labour markets have become more rigid Institutions did change 73
Explanation 1: Skilled Biased Technological Change Reminder: in chapter 2, we have seen how SBTC can explain the US path of wage inequality between skill groups Given globalization, all OECD countries face the same type of aggregate shocks SBTC Can SBTC be responsible for the rise in European unemployment? 74 Technological change and labour market institutions In Europe, wage rigidities work against SBTC, which tends to increase wage inequality Wage inequality does not increase, but unemployment rises If prices do not change, quantities must be modified 75
Some theory Reminder: In Chapter 2, we divided the labour force into educated (h) and uneducated individuals (l) p is the share of educated, while 1-p is the share of uneducated Relative supply of educated is worth l s = p/(1-p) w h is the wage of educated workers; w l is the wage of uneducated workers The schooling premium is the ratio ω = w h /w l 76 Some theory The relative skilled labour demand decreases with the skill premium Following SBTC, the relative labour demand curve moves rightward 77
Reminder: flexible wages ω ω 3 ω 1 ω 2 Sup Demand l 1 l 2 relative labor Relative demand increases more than relative supply, and the schooling premium rises 78 Some theory In Europe, wages are far from flexible: institutions such as unions and the minimum wage lower wage inequality A simple way to analyse the impact of such wage rigidity is to assume the wage premium remained fixed at its 1960 s level. Hence, wage inequality cannot increase as in the US. Then, what is the likely impact of technological change? 79
SBTC and wage rigidity The wage premium ω 1 stays at its previous equilibrium level ω Relative employment is then l 3, larger than competitive relative employment l 2 ω 1 rationing l 1 l 2 l 3 l How can the skilled and unskilled labour markets adjust? Unskilled employment becomes lower than the unskilled labour supply: there is unskilled unemployment The relative wage does not change, and relative employment varies 80 SBTC and wage rigidity If the relative wage is held fixed to the level reached before SBTC, relative employment cannot be equal to l 2 Relative employment is given by the relative labour demand curve. The ratio of employment ratios is worth: l 3 > l 2 81
SBTC and wage rigidity Finding the unskilled unemployment rate We knwo that l 3 is actual relative employment, while l 2 is the relative labour of skilled workers. Let u l be unskilled unemployment rate. We have Hence, p 2 ( 1 p2)( 1 u l) ul l2 = 1 l 3 2 = l3 l = l 1 u l 3 82 Explanation 2: Labour market institutions have changed Labour market institutions have changed during the past decades, and more especially two of them Unemployment compensation has become more generous Taxes on have increased, creating a fiscal wedge 83
Unemployment compensation Unemployment insurance provides higher income in many European countries than in the US. To appreciate the generosity of unemployment insurance, consider the following replacement rate unemployment benefit Replacemen t rate= previous wage As you know from Section 2, the generosity of unemployment compensation unfortunately alters job search: search effort declines and reservation wage increases 84 Average replacement rate and unemployment in Europe 10.6 Unemployment rate 8 6 4 2 1960 1970 1980 1990 2000.5.4.3.2 Average replacement rate Unemployment rate Replacement rate 85
Fiscal wedge How to finance the welfare state and social insurance in European countries? As lump-sum taxes do not exist, income and/or consumption are taxed Two main sources of income (neglecting land income, ie rents): returns to capital and returns to labour As capital is highly mobile while labour is not, most of the taxes are levied on wages The level and composition of labour and consumption taxation generates a fiscal wedge between the labour cost (paid by the employer) and the purchasing power (of the employee) 86 Anatomy of the fiscal wedge Labour cost = Wage + Payroll Taxes Real labour cost = Labour cost/(producer price) Net labour income = Wage Income Tax Purchasing power = Net labour income/(cons price) Consumer price = Producer Price + VAT Fiscal wedge = Purchasing Real labour power cost IT + VAT + PT Labour cost 87
Impact of the fiscal wedge Wage claims are based on the purchasing power, while the labour demand depends on the real labour cost Rising fiscal wedge cannot lower wage claims since workers do not directly benefit from the product of their taxes Hence, labour costs increase and unemployment rises 88 10 «Fiscal wedge» and European unemployment rate.35 Average unemployment rate 8 6 4.3.25 Average fiscal wedge 2.2 1960 1970 1980 1990 2000 Unemployment rate Fiscal wedge 89
Oil price from 1970 to 2001 Oil price 90 80 70 60 50 40 30 20 10 0 1970 1975 1980 1985 1990 1995 2000 nominal in 2001 dollars (real) 90 Oil shocks in the 1970s The full story Short run increase in unemployment rate, followed by an increase in labour taxation (welfare costs must be financed) an increase in replacement rate (political argument: unemployed have more political power when they get more numerous) the initial shock is prolongated The economy is attracted by a new long run equilibrium with high unemployment rate, generous unemployment compensation and high fiscal wedge 91
Hysteresis in equilibrium unemployment u 2 u 1 Up to t 0, the unemployment rate fluctuates around u 1 In t 0, a new recession starts (oil shock). This recession is stronger than the others When the shock stops in t 1, it s too late: the economy is now attracted by the high unemployment rate u 2 t 0 t 1 time 92 10 «Fiscal wedge» and European unemployment rate.35 Average unemployment rate 8 6 4.3.25 Average fiscal wedge 2.2 1960 1970 1980 1990 2000 Unemployment rate Fiscal wedge 93