The Ins and Outs of UK Unemployment

Transcription

1 Universiy of Warwick Special Session on Labour Marke Dynamics, March

2 Aim: To undersand wha drives changes in UK unemploymen. s i primarily job loss or job nding?

3 Aim: To undersand wha drives changes in UK unemploymen. s i primarily job loss or job nding? US: Shimer (2005, 2007); Hall (2005): mainly ou ows. Subsequen work suggess ha in ows may have a larger role.

4 Aim: To undersand wha drives changes in UK unemploymen. s i primarily job loss or job nding? US: Shimer (2005, 2007); Hall (2005): mainly ou ows. Subsequen work suggess ha in ows may have a larger role. Perongolo and Pissarides (2008): roughly spli. UK Labour Force Survey Excluding periods where unemploymen deviaes from seady sae.

5 Aim: To undersand wha drives changes in UK unemploymen. s i primarily job loss or job nding? US: Shimer (2005, 2007); Hall (2005): mainly ou ows. Subsequen work suggess ha in ows may have a larger role. Perongolo and Pissarides (2008): roughly spli. UK Labour Force Survey Excluding periods where unemploymen deviaes from seady sae. Burgess and Turon (2005): mainly in ows. UK Claiman Coun.

6 Conribuions hrough daa and mehod. Firs use of BHPS daa. An annual survey, bu hrough job hisories a monhly calendar is creaed Weigh so sample is represenaive of Briish populaion. ssues: Time aggregaion. Recall error. Misclassi caion error. Margin error.

7 Conribuions hrough daa and mehod. Firs use of BHPS daa. An annual survey, bu hrough job hisories a monhly calendar is creaed Weigh so sample is represenaive of Briish populaion. ssues: Time aggregaion. Recall error. Misclassi caion error. Margin error. Develop mehod o analyse changes in acual, raher han seady sae, unemploymen rae.

8 Conribuions hrough daa and mehod. Firs use of BHPS daa. An annual survey, bu hrough job hisories a monhly calendar is creaed Weigh so sample is represenaive of Briish populaion. ssues: Time aggregaion. Recall error. Misclassi caion error. Margin error. Develop mehod o analyse changes in acual, raher han seady sae, unemploymen rae. mporan because UK ransiion raes are low, so unemploymen deviaes from seady sae.

9 Conribuions hrough daa and mehod. Firs use of BHPS daa. An annual survey, bu hrough job hisories a monhly calendar is creaed Weigh so sample is represenaive of Briish populaion. ssues: Time aggregaion. Recall error. Misclassi caion error. Margin error. Develop mehod o analyse changes in acual, raher han seady sae, unemploymen rae. mporan because UK ransiion raes are low, so unemploymen deviaes from seady sae. Allow relaive in uence of job loss and job nding o vary over ime.

10 UK ransiion raes are low, relaive o US. Transiion rae UK US separaion E! U 0.5% 2% in ow from inaciviy! U 0.2% 3.5% job nding U! E 9% 32% ou ow o inaciviy U! 2% 33% Sources: UK is BHPS 1988q4-2007q2. US gures are Shimer (2007) CPS gross ows daa over he same period.

11 q1 1993q3 1998q1 2002q3 2007q1 Job finding rae (lef scale) BHPS acual unemploymen rae (lef scale) Separaion rae (righ scale) Quarerly averages of monhly raes. Source: BHPS.

12 Aim: decomposiion u = C in ow + C ou ow Conribuions of he various ows o he dynamics of he acual unemploymen rae are calculaed allowing pas ow raes o a ec curren unemploymen.

13 Aim: decomposiion u = C in ow + C ou ow Conribuions of he various ows o he dynamics of he acual unemploymen rae are calculaed allowing pas ow raes o a ec curren unemploymen. The low unemploymen exi rae means ha someone who loses heir job in he UK remains unemployed for abou 8 monhs, on average (a half life of around 6 monhs).

14 Aim: decomposiion u = C in ow + C ou ow Conribuions of he various ows o he dynamics of he acual unemploymen rae are calculaed allowing pas ow raes o a ec curren unemploymen. The low unemploymen exi rae means ha someone who loses heir job in he UK remains unemployed for abou 8 monhs, on average (a half life of around 6 monhs). Hall (2005): if urnover dynamics are relevan, seady sae unemploymen should lead acual unemploymen.

15

16

17 u = C in ow Aim and conribuions + C ou ow Acual unemploymen variaion depends on changes in curren ransiion raes (conribue o change in curren seady sae unemploymen rae, u ); changes in pas ransiion raes (capured recursively by u 1 ).

18 u = C in ow Aim and conribuions + C ou ow Acual unemploymen variaion depends on changes in curren ransiion raes (conribue o change in curren seady sae unemploymen rae, u ); changes in pas ransiion raes (capured recursively by u 1 ). u u 1 1 r u 1 r 1 r where r = 1 u = s +f s s +f s is he in ow rae f is he ou ow rae.

19 u = C in ow Aim and conribuions + C ou ow Acual unemploymen variaion depends on changes in curren ransiion raes (conribue o change in curren seady sae unemploymen rae, u ); changes in pas ransiion raes (capured recursively by u 1 ). C in ow C ou ow C in ow C ou ow 1 1 r + C in ow 1 1 r 1 r. 1 r + C ou ow 1 r 1 r.

20 u = C in ow Aim and conribuions + C ou ow Acual unemploymen variaion depends on changes in curren ransiion raes (conribue o change in curren seady sae unemploymen rae, u ); changes in pas ransiion raes (capured recursively by u 1 ). Perongolo and Pissarides (2008) showed for seady sae u: u = (1 u ) u 1 s s 1 {z } C in ow u (1 u 1 ) f. f {z 1 } C ou ow Can also break down overall ou ow and in ow rae conribuions o see he paricular conribuions of job nding and separaion raes.

21 A useful summary saisic of each componen s relaive imporance is is overall conribuion o he variance of seady sae unemploymen. These variance conribuions mirror nance beas. β ou ow = cov u, C ou ow var ( u ) Could summarise he conribuions wih a single number, averaging over ime.

22 A useful summary saisic of each componen s relaive imporance is is overall conribuion o he variance of seady sae unemploymen. These variance conribuions mirror nance beas. β ou ow = cov u, C ou ow var ( u ) Could summarise he conribuions wih a single number, averaging over ime. Or could allow for ime-varying beas.

23 Bea Bea in ow rae 0.61 separaion rae 0.49 in ow rae via inaciviy 0.12 ou ow rae 0.36 job nding rae 0.28 ou ow rae via inaciviy 0.08 Conribuions o unemploymen variance Noes. Bea is he conribuion o he variance of he acual unemploymen rae. Based on 12-monh moving averages of ransiion raes. Source: BHPS 1988q4-2007q2.

24 Variance conribuions of changes in ou ow and in ow raes o acual unemploymen rae dynamics (rolling 5-year beas)

25 Variance conribuions of changes in job nding, separaion and ransiion raes via inaciviy o acual unemploymen rae dynamics (rolling 5-year beas)

26 n ows and ou ows boh play a role.

27 n ows and ou ows boh play a role. The separaion rae seems o have been paricularly in uenial in raising unemploymen in he 1990s recession and in subsequen upicks.

28 n ows and ou ows boh play a role. The separaion rae seems o have been paricularly in uenial in raising unemploymen in he 1990s recession and in subsequen upicks. Conrary o Shimer-Hall claims.

29 n ows and ou ows boh play a role. The separaion rae seems o have been paricularly in uenial in raising unemploymen in he 1990s recession and in subsequen upicks. Conrary o Shimer-Hall claims. Bu in imes of moderaion, he job nding rae plays he bigges role. And policy can help boos is conribuion.

30 n ows and ou ows boh play a role. The separaion rae seems o have been paricularly in uenial in raising unemploymen in he 1990s recession and in subsequen upicks. Conrary o Shimer-Hall claims. Bu in imes of moderaion, he job nding rae plays he bigges role. And policy can help boos is conribuion. Conrary o Burgess-Turon (2005) conclusion.

31

32 Appendix Overall conribuion Curren conribuion in ow rae ou ow rae Comparison of curren and overall conribuions o unemploymen variance Noes. Bea is he conribuion o he variance of he acual unemploymen rae..the curren conribuion is he e ec of curren changes in ransiion raes o acual unemploymen dynamics. Based on 12-monh moving averages of ransiion raes. Source: BHPS 1988q4-2007q2.