Housing and labour dynamics

Transcription

1 Housing and labour dynamics How do house prices affect unemployment? Ekkehard Ernst 1 - Faten Saliba 2 1 Employment Trends Unit International Labour Organization Geneva ernste@ilo.org 2 saliba@ilo.org Project LINK Oct 23rd, 213 E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

2 Housing, asset prices and macro Housing market dynamics Recent housing price dynamics Price-income ratios are above historical averages for many countries E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

3 Housing, asset prices and macro Housing market dynamics Recent housing price dynamics Price-income ratios are above historical averages for many countries E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

4 Housing, asset prices and macro Housing vs. asset prices Comovement in housing and asset prices House and asset prices show co-movement over the long-run in the US E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

5 Housing, asset prices and macro Housing vs. asset prices Implication for labour markets (US) What is the proper lag? Employment growth (HP filtered) q1 198q1 1985q1 199q1 1995q1 2q1 25q1 21q1 215q Residential property prices (HP filtered) Employment growth Residential property prices (lagged by 4 years) Residential property prices E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

6 Overview Housing, asset prices and macro Overview of today s talk 1 Modelling housing markets Research strategy What drives labour market dynamics? 2 Housing markets and macro-labour Modelling and estimation strategy A housing-augmented double Phillips curve 3 Estimation and model dynamics Data and methodology Estimating the labour flow macro-model Model dynamics under shocks 4 Shock transmission under different policy settings 5 Concluding remarks E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

7 Modelling housing markets Research strategy Transmission mechanisms of housing prices Housing prices and the wealth effect Housing prices increase household wealth Consumption out of wealth leads to pro-cyclical business cycle movements Housing investment and geographical mobility Housing investment reduces geographical mobility Limits geographical and sectoral labour market adjustments House prices, wages and competitiveness House prices push wages and capital costs up Reduces firms competitiveness E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

8 Modelling housing markets Research strategy Model ideas I Flow model of the labour market Empirical formulation of standard matching model (e.g. Carlsson et al., 26) Full and separate account of unemployment in- and outflows Augmented with housing market channel Housing investment model Housing investment affects aggregate demand through a wealth effect......hiring via wage effects,......and competitiveness through productivity effects (e.g. Askenazy, 213) E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

9 Model ideas II Modelling housing markets Research strategy Wage-price dynamics Double Phillips curve (e.g. Flaschel et al. 1997; Erceg et al. 2): Reduced-form wage bargaining curve Hybrid Phillips curve Policy scenarios Short- vs. long-term effects of housing price dynamics Labour market policies Macro-prudential policies E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

10 Modelling housing markets An overview of the model flows I What drives labour market dynamics? Decomposing unemployment dynamics into......labour force growth and... U t = L t ET t = IN t OUT t L t = α 3 + β 31 L t 1 + β 32 u t 1 + β 33 Policy t...employment growth (i.e. the net effect of job creation and destruction) ET t = JobCreation t JobDestruction t E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

11 Modelling housing markets What drives labour market dynamics? An overview of the model flows II Job creation with housing investment JobCreation t = β 11 ET t 1 + β 12 w t + β 13 AD t + β 14 r t + β 15 Inv t + β 16 HousingInvestment t 1 Job destruction with housing investment JobDestruction t = β 21 TFP t + β 22 r t + β 23 REER t + β 24 AD t + β 25 w t + β 26 HousingInvestment t 1 Wage determination with housing investment: w t = α 4 + β 41 K t + β 42 Tax t + β 43 u t 1 + β 44 HousingInvestment t E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

12 Modelling housing markets What drives labour market dynamics? Putting the pieces together Substituting the flow equations: OUT t = JobCreation t IN t = JobDestruction t + L t 1 Hence: OUT t = β 11 OUT t 1 + β 12 Xt JobCreation + β 14 ET t 1 IN t = β 21 IN t 1 + β 22 Xt JobDestruction + β 24 L t 1 E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

13 Housing markets and macro-labour Modelling methodology I Modelling and estimation strategy Step-by-step estimation Step 1: Identify base-line equations Macro variables to affect unemployment flows Reduced-form panel estimates System-GMM used to control for endogeneity Results published in Ernst (211) Step 2: Identify relevant housing market interactions Labour flow model refers to aggregate demand, but separates out housing investment Two separate sectors with capital accumulation: Housing Non-housing Housing investment also affects productivity and wage growth E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

14 Housing markets and macro-labour Modelling methodology II Modelling and estimation strategy Step 3: Estimate macro model Introduce macro-economic closure: Modified Euler equation Introduce endogenous policy rules Estimate using GMM Step 4: Simulate model and reform scenarios Model simulation and shock analysis using Dynare Reform scenarios through parametric change Analysis of shock transmission: Productivity shocks Housing investment shock Asset price shock (share prices) Combined shocks Analyse impact on unemployment dynamics E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

15 Housing markets and macro-labour A housing-augmented double Phillips curve Housing markets and labour flows I Two-sector approach (a la Iacoviello, 21) Gross-fixed capital formation depends on profit outlook, public investment and real long-term interest rates: K t = K t 1 + F t 1 + G I t 1 + LP t 1 + r L t 1 + K Housing t Housing investment depends on demographics and real interest rates: Kt Housing = K Housing t 1 + rt 1 L + Pop t 1 Low productivity growth in construction Housing investment is less productive than in other sectors Average productivity declines with housing investment E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

16 Housing markets and macro-labour A housing-augmented double Phillips curve Housing markets and labour flows II Yield curve Wedge between long- and short-term interest rates Financial accelerator effect on long-term interest rates (Phelps, 1994) Short-term rates determined by household expectations, output gap and policy interventions Long-term rates with persistence determined by: Share prices Net government lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

17 Housing markets and macro-labour A Housing-price Phillips curve A housing-augmented double Phillips curve Price inflation, including imported inflation π t = π t 1 + Eπ t+1 + REER t 1 + Gap t 1 Wage inflation, includes a housing element w t = w t 1 + ET t 1 + π t 1 + Eπ t+1 + TFP t 1 + I Housing t 1 Output gap dynamics Gap t = w t + OUT t 1 + IN t 1 + GovCons t 1 E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

18 Estimation and model dynamics Data and methodology A word on the data and methodology Unemployment flows come from Elsby et al. (28) Constructed on the basis of information regarding unemployment duration at different duration lengths Complemented by similar information for more years and other countries to improve coverage Extended coverage possible using imputation methods with broadly similar results Information on share price dynamics is based on OECD share price index (OECD Main Economic Indicators) deflated by CPI Housing investment taken from OECD Economic Outlook database Other Macro indicators also come from the OECD Economic Outlook database Fixed effects have been accounted for through de-meaning: dx it = X it X i + X E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

19 Estimation and model dynamics Estimating the labour flow macro-model Estimation results: Labour block INt 1 LFPRt 1 Prodt 1 RIRSt 1 TaxIndt 1 Gapt Wagest 1 It Housing (1) INt.586*** *** -5.61***.6*** ***.38** *** (.26) (2.4) (.495) (.1) (1.53) (.3) (.142) (.489) OUTt 1 ETRt UCCt Wagest INVt Gapt It Housing (2) OUTt.582*** 1.36*** -.4* -1.7*** 3.71***.28*** 2.6*** (.28) (.218) (.2) (.325) (.764) (.3) (.716) OUTt INt (3) ETt.15*** -.19*** (.2) (.4) Wagest 1 πt 1 E {πt+1} Prodt ETt 1 It Housing (4) Wagest.656***.123**.48*.487***.255***.284*** (.4) (.63) (.27) (.82) (.42) (.15) Prodt 1 TFPt 1 It Housing (5) Prodt.714***.37*** -.11** (.91) (.7) (.5) E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

20 Estimation and model dynamics Estimating the labour flow macro-model Estimation results: Housing investment RIRLt 1 GAPt LFt 1 (6) I Housing t -.2**.2***.44*** (.1) (.) (.124) E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

21 Estimation and model dynamics Estimating the labour flow macro-model Estimation results: Macro block E {RIRSt+1} GAPt NLGQt πt 1 (7) RIRSt 1.157***.94*** -.116*** *** (.18) (.21) (.4) (1.384) RSharet RIRLt 1 RIRSt (8) RIRLt.7***.419**.493*** (.186) (.12) (.12) RSharet 1 GovInvt 1 Prodt 1 RIRLt 1 ETt 1 (9) INVt.4* 2.687***.678*** -.1***.547*** (.2) (.921) (.116) (.) (.34) OUTt 1 INt 1 Wagest GovConst 1 (1) GAPt 1.9*** -4.39*** 4.971*** 8.62*** (.131) (.278) (1.357) (2.698) πt 1 E {πt+1} ToTt 1 Wagest 1 (11) πt.449***.533*** -.49***.41 (.18) (.28) (.14) (.29) E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

22 Estimation and model dynamics Estimating the labour flow macro-model Estimation results: Fiscal block GovConst 1 ETt 1 (12) GovConst.973***.25* (.25) (.15) GovInvt 1 ETt 1 (13) GovInvt.959***.28*** (.42) (.7) OUTt 1 INt 1 (14) Taxt.1***.3 (.1) (.3) GovConst GovInvt Taxt (15) NLGQt *** *** *** (4.53) (9.23) (3.129) E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

23 Estimation and model dynamics Model dynamics under shocks Unemployment and housing price shocks An increase in housing investment raises unemployment through the competitiveness channel....1 Unemployment rate.2 Output gap 2 x 1 3 Investment Housing investment 4 x CPI inflation Labour x productivity x 1 3 Real wages Real long term interest rates Net government lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

24 Estimation and model dynamics Model dynamics under shocks Share prices and labour markets...whereas stronger asset prices improve the outlook for labour markets... Unemployment x rate x 1 4 Output gap 6 x 1 5 Investment Housing x investment x CPI inflation Labour 1 x productivity x Real wages Real long term interest rate Net government 4 x lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

25 Estimation and model dynamics Model dynamics under shocks Asset price bubble and labour markets...a simultaneous share-price-cum-housing boom benefits the labour market only temporarily.2 Unemployment rate.5 Output gap 4 x 1 3 Investment Housing investment CPI inflation Labour productivity Real wages 2 4 Long term interest rates Net government lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

26 Shock transmission under different policy settings Policies to deal with asset bubbles What type of policies to prevent asset price booms? Labour market policies Reduce wage pass-through of housing prices to avoid loss of competitiveness Macro-prudential regulation: Policy rate needs to react strongly to changes in asset prices Simulations through parametric changes Simulation 1: Reduction in elasticity of wages wrt to housing investment Simulation 2: Integration of asset prices in the short-term interest rate equation E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

27 Shock transmission under different policy settings Change labour market institutions Lower wage pass-through of housing boom.2 Unemployment rate.2 Output gap 5 x 1 3 Investment Housing investment CPI inflation Labour productivity Real wages Long term real interest rates Net government lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

28 Shock transmission under different policy settings Macro-prudential regulation Changes in the reactivity of the policy rate to asset prices.5 Unemployment rate 1 Output gap.2 Investment Housing investment CPI inflation Labour productivity Real wages Long term real interest rates Net government lending E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

29 Concluding remarks Lessons learned and outlook What have we learned so far? Linear structural model allows full estimation of all parameters Pass-through play important role for labour flows and unemployment Allows detailed analysis of transmission mechanisms In particular, transmission of housing and asset prices can be easily assessed Next steps Labour flow data available for 8 countries Develop open economy and global model (regional blocks) Allow for international spill-overs Integrate housing prices directy, rather than housing investment E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

30 Bibliography Bibliography I Askenazy, P Capital Prices and Eurozone Competitiveness Differentials, IZA Discussion Paper, No (Available at: Carlsson, M.; Eriksson, S.; Gottfries, N. 26. Testing Theories of Job Creation: Does Supply Create Its Own Demand? Discussion Paper, No. 224 (Bonn: Institute for the Study of Labor (IZA)). Erceg, C. J.; Henderson, D. W.; Levin, A. T. 2. "Optimal monetary policy with staggered wage and price contracts", Journal of Monetary Economics, Vol. 46, No. 2, pp Ernst, E Determinants of unemployment flows. Labour market institutions and macroeconomic policies. Discussion paper, No. 29 (Geneva: International Institute for Labour Studies), available at: Ernst, E.; Rani, U Understanding unemployment flows, Oxford Review of Economic Policy, Volume 27, No. 2, pp E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31

31 Bibliography Bibliography II Flaschel, P.; Franke, R.; Semmler, W Dynamic macroeconomics. Instability, fluctuations, and growth in monetary economies. (Cambridge, MA.: MIT Press). Iacoviello, M. and Neri, S. 21. "Housing market spillovers: Evidence from an estimated DSGE model, American Economic Journal: Macroeconomics, vol. 2, pp (Available at: Phelps, E Structural Slumps. (Cambridge, MA.: Harvard University Press). E. Ernst & F. Saliba (ILO) ILO Modelling New York, / 31