Beveridge Curve Shifts across Countries since the Great Recession

Transcription

1 13TH JACQUES POLAK ANNUAL RESEARCH CONFERENCE NOVEMBER 8 9, 2012 Beveridge Curve Shifts across Countries since the Great Recession Bart Hobijn Federal Reserve Bank of San Francisco Ayşegül Şahin Federal Reserve Bank of New York Paper presented at the 13th Jacques Polak Annual Research Conference Hosted by the International Monetary Fund Washington, DC November 8 9, 2012 The views expressed in this paper are those of the author(s) only, and the presence of them, or of links to them, on the IMF website does not imply that the IMF, its Executive Board, or its management endorses or shares the views expressed in the paper.

2 Bart Hobijn (FRBSF, VU Amsterdam, and TI) and Ayşegül Şahin (FRBNY) BEVERIDGE CURVE SHIFTS ACROSS COUNTRIES SINCE THE START OF THE GREAT RECESSION Disclaimer: The views expressed in this presentation are those of the authors and do not necessarily reflect those of the Federal Reserve Bank of San Francisco, the Federal Reserve Bank of New York, the Federal Reserve System as a whole, or any other institutions that they are affiliated with. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 1

3 Stable pre-recession Beveridge curve Actual Beveridge Curve Monthly observations; seasonally adjusted Job openings rate 5% before 2007 recession 4% 3% 2% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% Unemployment rate Source: JOLTS, CPS, and authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 2

4 fell apart since Great Recession Actual Beveridge Curve Monthly observations; seasonally adjusted Job openings rate 5% before 2007 recession 4% 3% Jun-12 since 2007 recession 2% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% Unemployment rate Source: JOLTS, CPS, and authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 3

5 2.8 percentage points rightward shift Beveridge Curve and shift Monthly observations Job openings rate 5% before 2007 recession 4% Gap: 2.8% Jun-12 3% since 2007 recession 2% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% Unemployment rate Source: JOLTS, CPS, and authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 4

6 Is this shift unprecedented? Use cross-country evidence to put recent shift in U.S. Beveridge curve in global and historical perspective Hobijn and Sahin (11/08/2012) Beveridge curve shifts 5

7 No, it is not Use cross-country evidence to put recent shift in U.S. Beveridge curve in global and historical perspective Current U.S. Beveridge curve shift fits standard profile (across countries and over time) Deep recession leads to large displacement of workers, resulting in mismatch. Policy response of increased UI generosity to alleviate burden of displacement. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 6

8 Part 1: BEVERIDGE CURVE AS LABOR TURNOVER STEADY-STATE CONDITION Barnichon, Elsby, Hobijn, and Şahin (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 7

9 Unemployment and labor turnover Unemployment rate constant when Hires rate = Separations rate + Entrants into labor force - Exits out of the labor force Hobijn and Sahin (11/08/2012) Beveridge curve shifts 8

10 Unemployment and labor turnover Unemployment rate constant when Hires rate = Separations rate + + Entrants into labor force -- Exits out of of the labor force Hobijn and Sahin (11/08/2012) Beveridge curve shifts 9

11 Beveridge curve as steady-state condition Unemployment rate constant when Hires rate = Separations rate Beveridge curve: For every unemployment rate, find the vacancy rate at which above equation holds. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 10

12 Fitted Pre-Recession U.S. Beveridge curve Actual and fitted Beveridge Curve Monthly observations; pre-2007-recession fit Job openings rate 5% Fitted before 2007 recession 4% Gap: 2.8% Jun-12 3% since 2007 recession 2% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% Unemployment rate Source: JOLTS, CPS, and authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 11

13 Decline in match efficiency Actual and Fitted Vacancy Yield Monthly data; seasonally adjusted; regression ends 11/2007 End of regression sample hires per vacancy 2.5 Fitted 2 Actual Source: BLS and authors' calculations 0 Borowczyk-Martins et al. (2011), Davis, Faberman, and Haltiwanger (2012), and Sedláček (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 12

14 Workers stopped quitting jobs Actual and Fitted Quits Rate (quits per employee) Monthly data; seasonally adjusted; regression ends 11/2007 End of regression sample share of employed Fitted 0.02 Actual Source: BLS and authors' calculations 0 Lazear and Spletzer (2012), Saint-Paul (1995) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 13

15 Effect on position of Beveridge curve At same unemployment and vacancy rate Fewer people get hired. Fewer people quit their jobs. Hires effect dominates For a given level of unemployment, we need more vacancies to keep unemployment constant Causes a rightward (upward) shift in Beveridge curve Hobijn and Sahin (11/08/2012) Beveridge curve shifts 14

16 Part 2: CAUSES OF U.S. BEVERIDGE CURVE SHIFT Elsby, Hobijn, and Şahin (2010), Daly, Hobijn, Şahin, and Valletta (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 15

17 Three main sources of shift $ House lock UI extensions Mismatch Hobijn and Sahin (11/08/2012) Beveridge curve shifts 16

18 Least evidence for house lock $ House lock Unimportant Oswald (1997), Nickell et al. (2001) Schulhofer-Wohl (2010), Molloy, Smith, and Wozniak (2011), Valletta (2010), Sterk (2010), Karahan and Rhee (2012) UI extensions Mismatch Hobijn and Sahin (11/08/2012) Beveridge curve shifts 17

19 UI extensions important but temporary $ House lock Unimportant UI extensions Expected to expire Mismatch Oswald (1997), Nickell et al. (2001) Schulhofer-Wohl (2010), Molloy, Smith, and Wozniak (2011), Valletta (2010), Sterk (2010), Karahan and Rhee (2012) Aaronson, Mazumder, and Schecter (2010), Valletta and Kuang (2010), Fujita (2010), Farber and Valletta (2011), Rothstein (2011), Nakajima (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 18

20 Mismatch most substantial factor $ House lock Unimportant UI extensions Expected to expire Mismatch Largely transitory due to dynamics nature of U.S. market. Oswald (1997), Nickell et al. (2001) Schulhofer-Wohl (2010), Molloy, Smith, and Wozniak (2011), Valletta (2010), Sterk (2010), Karahan and Rhee (2012) Aaronson, Mazumder, and Schecter (2010), Valletta and Kuang (2010), Fujita (2010), Farber and Valletta (2011), Rothstein (2011), Nakajima (2012) Şahin, Song, Topa, and Violante (2011), Dickens and Triest (2012), Daly, Hobijn, Şahin and Valletta (2012), Foote (2012), Hobijn (2012), Lazear and Spletzer (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 19

21 Part 3: LABOR TURNOVER BEYOND THE U.S. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 20

22 No JOLTS data for other countries Most countries do not run a Job Openings and Labor Turnover Survey. No direct measure of hires and separations. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 21

23 Alternative way to impute labor turnover Most countries do not run a Job Openings and Labor Turnover Survey. No direct measure of hires and separations. Labor turnover can be imputed from data on employment by job tenure. Construct annual time-series of hires and separations using new method. OECD (2009) Cyclicality of turnover is what is important for our analysis Hobijn and Sahin (11/08/2012) Beveridge curve shifts 22

24 Main intuition of imputation method Employment and labor turnover in France 2006 Millions of persons employed Millions of persons Employed shorter than a year Employed longer than a year Source: OECD Shimer (2005,2012), Elsby, Hobijn, and Şahin (2010) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 23

25 Hires derived from newly employed Employment and labor turnover in France 2006 Millions of persons employed Millions of persons 30 Hires 3.8 Employed shorter than a year Employed longer than a year Source: OECD Shimer (2005,2012), Elsby, Hobijn, and Şahin (2010) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 24

26 Separations imputed from job leavers Employment and labor turnover in France 2006 Millions of persons employed Millions of persons 30 Separations 3.8 Employed shorter than a year Employed longer than a year Source: OECD Shimer (2005,2012), Elsby, Hobijn, and Şahin (2010) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 25

27 United States Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Similar cyclicality of turnover across countries Estimated Cyclicality of vacancy yield and separations rate Elasticity of labor turnover with respect to (u/v)-ratio JOLTS OECD Employment by job tenure data Elasticity Hires/vacancy Separations rate Source: JOLTS, OECD, and author's calculations Petrongolo and Pissarides (2001) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 26

28 Part 4: TWO PARTICULAR COUNTRIES: SPAIN AND GERMANY Hobijn and Sahin (11/08/2012) Beveridge curve shifts 27

29 Spain s slow dynamics Spain: Actual and steady-state unemployment rates Annual data Share of labor force 30% Actual 25% Steady State 20% 15% 10% 5% 0% Elsby, Hobijn, and Şahin (2010) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 28

30 result in loose Beveridge curve Actual and Fitted Beveridge Curve: Spain Annual data; fitted curved based on pre-2007 data on tenure. Vacancy rate (index avg=100) Non sample Sample Great Recession and after Fitted Unemployment rate Source: OECD, Nickell and Nunziata (2004), Eurostat (2012), and authors' calculations Hobijn and Şahin (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 29

31 Match efficiency decline in Spain Estimated Vacancy Yield: Spain annual data log hires per vacancy Actual Fitted Source: OECD and authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 30

32 Continued effect of reforms in Germany Actual and Fitted Beveridge Curve: Germany Annual data; fitted curved based on pre-2007 data on tenure. Vacancy rate (index avg=100) 300 Non sample Sample Great Recession and after Fitted Unemployment rate Source: OECD, Nickell and Nunziata (2004), Eurostat (2012), and authors' calculations 0 Hobijn and Sahin (11/08/2012) Beveridge curve shifts 31

33 Very few separations in Germany Estimated Separations Rate: Germany Annual observations Actual Fitted log separations over Source: OECD and authors' calculations -2.5 Burda and Hunt (2011) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 32

34 Part 5: SHIFTS SINCE 2007 ACROSS COUNTRIES Hobijn and Sahin (11/08/2012) Beveridge curve shifts 33

35 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Shifts in Beveridge curves since 2007 Deviations from fitted Beveridge Curve: Average percentage point deviation in the unemployment rate, Percent Source: Authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 34

36 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Portugal, Spain, Sweden, U.K. stand out Deviations from fitted Beveridge Curve: Average percentage point deviation in the unemployment rate, Percent Source: Authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 35

37 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Shifts reflected in labor turnover Shortfall/excess in labor turnover 2010 percent deviation from fitted value Percent Hires/vacancy Separations rate Source: Authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 36

38 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Shifts reflected in labor turnover Shortfall/excess in labor turnover 2010 percent deviation from fitted value Percent Hires/vacancy Separations rate Source: Authors' calculations Hobijn and Sahin (11/08/2012) Beveridge curve shifts 37

39 Part 6: CAUSES AND HISTORICAL PERSPECTIVE Hobijn and Sahin (11/08/2012) Beveridge curve shifts 38

40 Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom United States U.S. and Sweden extended UI Duration of Unemployment Insurance Benefits Duration in months for 2007 and Months Source: OECD (2012) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 39

41 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom United States Mismatch/house lock is rest of story Percent 6 Change in construction employment and house prices change in share of construction employment and house prices Percent Construction employment (left axis) House prices (right axis) Source: OECD Hobijn and Sahin (11/08/2012) Beveridge curve shifts 40

42 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom United States Mismatch/house lock is rest of story Percent 6 Change in construction employment and house prices change in share of construction employment and house prices Percent Construction employment (left axis) House prices (right axis) Source: OECD Hobijn and Sahin (11/08/2012) Beveridge curve shifts 41

43 Australia Austria Belgium France Germany Japan Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom United States Shifts small compared to 80s Current and historical Beveridge curve shifts Shifts since 2007 and from '75-'85 to '91-'07 fitted curves Since s to '90s Percent Source: authors' calculations Blanchard and Wolfers (2000), Nickell et al. (2001) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 42

44 In case you just walked in Use cross-country evidence to put recent shift in U.S. Beveridge curve in global and historical perspective Current U.S. Beveridge curve shift fits standard profile (across countries and over time) Deep recession leads to large displacement of workers, resulting in mismatch. Policy response of increased UI generosity to alleviate burden of displacement. Hobijn and Sahin (11/08/2012) Beveridge curve shifts 43

45 U.K. experience a roadmap for U.S.? Beveridge Curve: United Kingdom Annual data Vacancy rate (index 100 mean) Unemployment rate Source: OECD and authors' calculations Pissarides (2006) Hobijn and Sahin (11/08/2012) Beveridge curve shifts 44