TFP Decline and Japanese Unemployment in the 1990s Julen Esteban-Pretel Ryo Nakajima Ryuichi Tanaka GRIPS Tokyo, June 27, 2008
Japan in the 1990s The performance of the Japanese economy in the 1990s was less than stellar Hayashi, F. and E. C. Prescott, 2002 1990s in Japan: a lost decade Review of Economic Dynamics 5 (1), 206-235. 2
The 1990s: Output Average annual growth rate of GNP per capita 1991-2000 Japan: 0.5% U.S.: 2.6% 0.2 Log of Real GNP (1990=1) 0-0.2-0.4 U.S. Japan -0.6-0.8 1960/Q1 1964/Q1 1968/Q1 1972/Q1 1976/Q1 1980/Q1 1984/Q1 1988/Q1 1992/Q1 1996/Q1 2000/Q1 3
The 1990s: Detrended Output 105 Detrended Real GNP 100 95 90 85 80 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1980s: Economy grows more than the benchmark 2%. 1990s: Trend reversal. 2003 is 90% of that of 1990 had it grown at a 2% rate per year. 4
The 1990s: Unemployment 6.0 Unemployment Rate 4.5 Percent 3.0 1.5 0 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 1980s: 2.5% unemployment rate. 1990s: Big increase in unemployment. 5.4% unemp. rate. in 2002. 5
Forces Behind Unemployment Increase Creation ut nt Destruction 6
Probability of Finding Job 45 Quarterly Prob. of Finding Job 40 Percent 35 30 25 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1990: 40% of unemployed workers found job within 1 quarter. 2000: Probability reduced to 30%. 7
Probability of Losing Job 2.0 Quarterly Prob. of Losing Job 1.5 Percent 1.0 0.5 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1990: 0.8% of workers lost job every quarter. 2000: Probability increased to 1.8%. 8
Number of Jobs Created and Destroyed 2.0 Jobs Created and Destroyed as Fraction of Labor Force Percent of Labor Force 1.5 1.0 Destruction Creation 0.5 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Over the 1990s creation and destruction increased. Destruction was higher than creation. 9
Lost Decade Labor Market Facts Detrended output decreased. Unemployment increased. Increase in destruction. Increase in creation, but lower than destruction. Hours of work decreased. Detrended wages decreased. 10
What Drove the Lost Decade? Possible explanations: Krugman (1998), Caballero et al. (2007) Bad fiscal policy. Liquidity trap. Depressed investment (over-investment during bubble period). Credit constraints. Hayashi and Prescott (2002) TFP slowdown. 11
TFP 0.012 Total Factor Productivity 0.010 0.008 0.006 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 A t = ϕ t z t where z t = z t 1 e γ ϕ t = ϕ 1 ρ ϕ ρ t 1 eɛ t 12
TFP Growth ( γ) 0.06 TFP Growth 0.04 0.02 0-0.02-0.04-0.06 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 1970-1990: 2.1% growth. 1991-2000: 0.2% growth. 1998-2003: 1.2% growth. 13
Detrended TFP (φ) 105 Detrended TFP 100 95 90 85 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Detrended TFP drops below 90% of the 1990 level. 14
This Paper Can the decline in detrended TFP generate the increase in unemployment observed in Japan over the 1990s? 15
What We Do We build: Neo-Classical growth model with search frictions in the labor market. Calibrate it to match the Japanese economy in 1990. Simulate it using TFP data and compare results with data. 16
Outline Model General structure Parameterization Main equations Results Conclusions 17
General Structure Cass-Koopmans type model with labor market search frictions. Two types of agents: Households: - Consume. - Save. - Supply labor to firms. Firms: - Hire labor and rent capital to produce output. - Atf(kt), where At : TFP. kt : Detrended capital per worker. Labor market: Search and Matching. Endogenous destruction. 18
Household s Problem Economy composed by a big family perfect self insurance. Firms are own by the Family. The Family chooses to max {C t+i, K t+i+1 } i=0 E t i=0 β i log (C t+i ) s.t FOC: C t+i + K t+i+1 = W t+i +Π t+i +(1 + r t+i δ) K t+i +(1 n t+i ) bz t+i { } C 1 t = βe t (1 + r t+1 δ) C 1 t+1 19
Labor Market Search and matching labor market. Unemployed workers and vacancies meet randomly according to CRS matching function m(ut,vt)= m(θt)ut, where θt = vt/ut. Prob. for an unemp. worker to match with a firm: qt = m(θt)/θt. Prob. for firm to match with unemp. worker: pt = m(θt). Production function: At kt α. At : TFP - At =φtzt, zt = zt-1e γ, ϕ t = ϕ 1 ρ ϕ ρ t 1 eɛ t - γ Growth rate of TFP, φ SS detrended TFP. Idiosyncratic cost, xt, - i.i.d with dist. fn: G : [0,xmax] [0,1]. Endogenous destruction: depends on value of xt. Free entry of firms. Nash bargaining for wages. 20
Problem of the Firm Value of a vacancy: V t = φz t + β [ q t max { } ] E t J t+1, V t+1 +(1 qt ) V t+1 x t+1 V t = φz t + E t β t q t J t+1 (x t+1 ) dg (x t+1 ) + (1 q t G (x t+1 )) V t+1 0 = φz t + E t β t q t 0 x t+1 J t+1 (x t+1 ) dg (x t+1 ) Value of filled job: 0 J t (x t )=A t f (k t ) r t z t k t x t z t w t (x t )+β t max { E t J t+1, V t+1 } J t (x t )=A t f (k t ) r t z t k t x t z t w t (x t )+E t β t J t+1 (x t+1 ) dg (x t+1 ) x t Optimal capital: r t = f (k t ) 0 21
Problem of the Worker Value of unemployment: U t = bz t + β t [p t max {E t N t+1, U t+1 } +(1 p t ) U t+1 ] x t+1 U t = bz t + E t β t p t N t+1 (x t+1 ) dg (x t+1 ) + (1 p t G (x t+1 )) U t+1 Value of employment: 0 N t (x t ) = w t (x t ) + βmax {E t N t+1, U t+1 } x t+1 N t (x t ) w t (x t )+E t β t N t+1 (x t+1 ) dg (x t+1 ) + (1 G (x t+1 )) U t+1 0 22
Wage, Surplus and Threshold Wages are determined as the Nash solution to a bargaining problem, where η is the bargaining power of the worker. This problem delivers: max (N t (x t ) U t ) η (J t (x t ) V t ) 1 η w t (x t ) N t (x t ) U t = ηs t (x t ) J t (x t ) = (1 η) S t (x t ) where S t (x t ) = J t (x t ) + N t (x t ) U t Wage: w t (x t )=η [A t f (k t ) z t r t k t x t z t + φz t θ t ] + (1 η) bz t Destruction threshold: S t (x t ) = 0 23
Flows and Timing t - 1 n t u t Production Matching t A t+1, x t+1 Destruction of jobs n t+1 u t+1 t + 1 [1 G (x t+1 )] ut p t G (x t+1 ) nt Flows: u t = [ 1 p t 1 G (x t ) ] u t 1 +[1 G (x t )] n t 1 1 = n t + u t 24
Aggregation Aggregate output: Y t = n t A t k α t Aggregate capital (capital market clearing condition): Economy resource constraint: K t = n t A t k t Y t +(1 n t ) bz t = C t + K t+1 (1 δ) K t + φz t v t + n t x t z t 25
Stationary Competitive Equilibrium A stationary competitive equilibrium in this economy is a set of: unemployment rate, employment rate and vacancies, ut, nt, vt; value for the firms and workers, Jt(xt), Nt(xt),Ut; wages, wt(xt); intermediate input cost threshold, x t ; interest rate, and detrended cons., output and capital (agg. and per worker), r t, C t, Ỹt, K t, y t, k t. which satisfy: flow equations of the labor market; optimal values for firms and workers, and optimal capital condition; Nash bargaining for wage determination; zero surplus condition; households optimization condition; economy resource constraint, capital market clearing condition. 26
Outline Model Parameterization Choose functional forms. Results Fix a set of parameters exogenously. Conclusions Calibrate other parameters to match Japanese data in 1990. 27
Functional Forms Matching function: m(u,v) = µ u ξ v 1-ξ Idiosyncratic productivity distribution: Exponential: - mean: µn 28
Exogenous Parameters A subset of parameters are fixed using values in other studies or in an ad-hoc manner. Exogenously fixed parameters Discount factor β = 0.98 Elasticity of matching with respect to unemp. ξ = 0.5 Bargaining power of the worker η = 0.5 Exponent of capital in prod. function α = 0.365 Capital depreciation δ = 0.022 29
Calibrated Parameters Calibration is done to match the Japanese economy in 1990. Calibrated parameters Parameters (4) Moments Matched (4) Source ϕ : Cost of posting vacancy Unemployment rate (1990) 2.1% b : Flow value of unemp. Prob. of leaving unemp. (1990) 0.42 µ : Scaling param in match. fn Ratio of unemp. benefit to output 0.4 Kuroda (2003) Shimer (2004) µn : Mean of G( ) Market tightness, θ, set to 1 Shimer (2004) 30
Outline Model Parameterization Results Feed detrended FTP from the data. Conclusions Terminal detrended TFP is 88% that of 1990. Compare the changes between steady states. Simulate path to new SS using perfect foresight shooting algorithm. 31
Detrended TFP 105 Detrended TFP 100 95 90 85 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1990: 100. 2003: 88. 32
Changes in SS of Output & Unemployment Unemployment: Data: 1990-2.1% 2003-5.2% Model: 1990-2.1% 2003-4.7% Output (detrended): Data: 1990-100 2003-87 Model: 1990-100 2003-83 33
Path of Unemployment to the new SS Unemployment 6.0 4.5 5.2% 4.7% Percent 3.0 2.1% 1.5 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Data Model 34
Path of Output to the new SS Detrended Output 105 100 100 95 90 90 85 83 80 75 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Data Model 35
Destruction of Jobs Total Jobs Destroyed per Quarter 2.0 1.7 1.8 Percent of Labor Force 1.4 1.1 1.3 0.8 0.8 0.5 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Data Model 36
Creation of Jobs Total Jobs Created per Quarter 1.5 1.4 Percent of Labor Force 1.3 1.1 0.9 0.88 1.25 0.7 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Data Model 37
Destruction vs Creation of Jobs Total Jobs Destroyed and Created per Quarter 1.5 1.3 Percent of Labor Force 1.1 0.9 0.7 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Destruction Creation 38
Detrended Wages Detrended Wages 105 100 100 95 90 85 80 81.7 81.5 75 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Data Model 39
Conclusions We build a Neo-Classical Growth Model with search frictions in the labor market. We show that decreases in detrended TFP can account for a big part of the behavior over the 1990s of: Output. Unemployment. Creation and destruction of jobs. Wages. Future work: Add hours of work. Add government expenditure. 40