Schweizerische Nationalbank · SIX:SNBN

Schweizerische Nationalbank — Swiss National Bank — Technical Presentation (May 2011)

32 slides from Schweizerische Nationalbank's technical presentation, May 15, 2011

  1. Slide 1: Introducing financial frictions in DSGE-CH

    Financial frictions in DSGE models Introducing financial frictions in DSGE-CH DSGE CH Jean-Marc Natal Financial Stability – Banking System Swiss National Bank

  2. Slide 2: Why should we care?

    Financial frictions in DSGE models Why should we care?  Since financial crisis, financial stability considerations take increasing importance in policy discussions at CBs and fiscal authorities  Questions:  Importance of asset prices for monetary policy?  Interactions btw bt monetary monetar and

  3. Slide 3: New macro models

    Financial frictions in DSGE models New macro models  B Boom in i literature li introducing i d i active i role l for b k (CS) in f banks i DSGE models: d l Gertler and Karadi, 2009, Gertler and Kiyotacki, 2010, Gerali et al., 2010, …  Focus on asymmetric information between banks and their credito

  4. Slide 4: Outline

    Financial frictions in DSGE models Outline 1. Benchmark DSGE-CH: performance and shortcomings 2. Introducing financial intermediation (CD): DSGE-CH II a. Structure b. How fin. frictions affect standard shock transmission ? c. What is the effect of financial shocks ? 3. Work in progress DSGE-CH III:

  5. Slide 5: 1. Benchmark DSGE-CH (2007)

    Financial frictions in DSGE models 1. Benchmark DSGE-CH (2007) ( )  Medium-sized ((60 eq.) q ) SOE-DSGE model of the Swiss economyy used for: 1. Forecasting (quarterly) since 2007Q1 2. Scenario analysis, “what if exchange rate…?” 3. Optimal policy analysis - Fixed vs floating e-rate - Inflation vs

  6. Slide 6: Decent overall (pseudo) forecasting performance …

    Financial frictions in DSGE models Decent overall (pseudo) forecasting performance … Forecast conditioned on: y*, p*, r*, poil 6

  7. Slide 7: … and surprisingly also during the crisis.

    Financial frictions in DSGE models … and surprisingly also during the crisis. Forecast conditioned on: y*, p*, r*, poil 7

  8. Slide 8: Comparing with housing/banking crisis of 90's

    Financial frictions in DSGE models Comparing with housing/banking crisis of 90’s Forecast conditioned on: y*, p*, r*, poil 8

  9. Slide 9: A puzzle?

    Financial frictions in DSGE models A puzzle? Source: M. Drehmann, December 2009 9

  10. Slide 10: 2009/2010: no severe disintermediation !

    Financial frictions in DSGE models 2009/2010: no severe disintermediation ! Source: M. Drehmann, December 2009 10

  11. Slide 11: Good luck? Good policy?

    Financial frictions in DSGE models Good luck? Good policy? Need a model to run counterfactuals 11

  12. Slide 12: Introducing financial frictions: strategy

    Financial frictions in DSGE models Introducing financial frictions: strategy  Two stages (away of representative agent) 1. DSGE-CH II (CD): asymmetric info problem between banks and entrepreneurs t BGG (1999) 2. DSGE-CH III (CS): asymmetric info problem between banks and banks’ share- banks share a

  13. Slide 13: 2. DSGE-CH II (2010 - …)

    Financial frictions in DSGE models 2. DSGE-CH II (2010 - …)  Moral hazard problem between banks and entrepreneurs (SOE version of BGG, 1999)  New agent: entrepreneurs has tendency to underreport profit  Moral hazard problem requires costly monitoring financed by premium on loans (EFP)  The more

  14. Slide 14: Small open economy model structure

    Small open economy Exog. domestic shocks (g,A,mon) Goods y=i+c+g market Exog. foreign shocks clearing (y,P,i,pe) II. Final goods y = f(e,Xd,Xm) production (non tradable) Xm Imports I. Intermediate goods d X= Xd+ Xf Exports production (tradable) g assets ((F)) Net foreign Intermediate K,e,L Firms

  15. Slide 15: Small open economy with entrepreneurs and banks

    Small open economy Exog. domestic shocks (g,A,mon, risk, equity price) y=i+c+g Exog. foreign shocks (y,P,i,pe) y = f(e,Xd,Xm) Xm Imports X= Xd+ Xf Exports g assets ((F)) Net foreign K,e,L Household Modified UIP (ds) (1+id )D Entrepreneur (1+ik )Bk Bank Entreprene r Entrepreneur Entrepreneur Risk

  16. Slide 16: Transmission of standard shocks

    Financial frictions in DSGE models Transmission of standard shocks 16

  17. Slide 17: Transmission of standard shocks

    Financial frictions in DSGE models Transmission of standard shocks 17

  18. Slide 18: Amplification of asset price shocks

    Financial frictions in DSGE models Amplification of asset price shocks 18

  19. Slide 19: Risk has 1st order effect on activity (Christiano et al. 2010)

    Financial frictions in DSGE models Risk has 1st order effect on activity (Christiano et al. 2010) 19

  20. Slide 20: In a nutshell

    Financial frictions in DSGE models In a nutshell  Standard model is a good approximation for transmission of macro- shocks. Its use in “normal times” should not be a cause for concern.  When asset prices are grossly out of line with fundamentals, the standard model is bad approximation.  DSGE-CH

  21. Slide 21: 3. DSGE-CH III (work in progress)

    Financial frictions in DSGE models 3. DSGE-CH III ((work in p progress) g )  Model is based on Gertler and Karadi (JME, 2010), Gertler and Kiyotacki (2010), Gerali et al. (JMCB, 2010), Bernanke et al. (1999) and Dib (2010).  Introduces risky banking: bankers may default/divert assets if the cost o

  22. Slide 22: Financial accelerator: DSGE-CH II

    Y=f(K,L,e) Financial accelerator: DSGE-CH II Firms L K I C Entrepreneurs Labor market (1+iE)BE Household (1+id)D Bank 22 (1+iE)/(1+id)= 1+EEFP(Entr. leverage)

  23. Slide 23: DSGE-CH III: double agency problem

    Y=f(K,L,e) Y=f(K,L DSGE-CH III: double agency problem Firms L K I C Entrepreneurs Labor market (1+iE)BE Lending bank Dividends Saving Household (1+id)D bank (1+ib)B 23 (1+iE)/(1+id) = [1+EEFP(Entr. leverage)]x[1+BEFP(Lending bank leverage)]

  24. Slide 24: Looking ahead

    Financial frictions in DSGE models Looking ahead  DSGE models are promising tools to assess the impact of financial shocks on the economy and evaluate alternative macroprudential/monetary policies  But: existing state of the art models do not explain:  i) endogenous d increases i in i leverage l

  25. Slide 25: Appendix

    Financial frictions in DSGE models Appendix 25

  26. Slide 26: CVAR

    Inflation GDP growth yoy — Actual Recursive out-of-sample forecasts — Actual Recursive out-of-sample forecasts 3.5 5.0 3.0 4.0 25 2.0 3.0 5 Ls — 20 r. 10 05 1.0 0.0 a 1.0 05 10 -10 2000 2001 2002 2003 2014 2005 2006 2007 2008 2009 2010 2011 Three-month Libor — Actual Recursive out-of-sample forecast

  27. Slide 27: LSEM

    Chart 8: Forecasting performance — rolling parameter estimates CPI inflation yoy in % — Actual — Forecasts 35 3.0 2.5 20 OS -10 10" 94 95 96 97 98 99 00 01 O2 03 OF 05 06 O7 O8 Three-month Libor in % — Actual — Forecasts 94°95 96 97 98 99 OO OF 02 03 OF OS 06 OF OB GDP growth gog in %, annualised —

  28. Slide 28: DSGE-CH

    Inflation Interest rate — Observed Forecasts — Observed Forecasts 7 12 DSGE-CH 85 90 95 00 os

  29. Slide 29: Table 4: RMSE for CVAR model

    Table 4: RMSE for CVAR model Model Variable Forecast horizon 1 2 3 4 8 12 CVAR in sample as O47 O51 O57 063 0.70 0.70 1982 Q1 to 2008 Q4 PM 051 055 O57 056 063 0.72 y 063 O69 O77 40.92 1.04 1.06 GVAR in sample 7 033 O57 0.79 O92 159 2.42 CVAR recursive 7 025 O39 052 O67 £40.85 0.68 1995 Q1 to 2008 Q

  30. Slide 30: LSEM – in sample:

    LSEM — in sample: Table 1: Forecasting performance - Parameters estimated on full sample Forecast horizon (in quarters) 1 2 3 4 5 6 7 8 9 10 11 12 #13 14 #15) «16 CPI inflation Root-mean-squared errors (in percentage points) GDP growth 3-month Libor 0.34 0.51 0.65 0.72 0.67 0.65 0.62 0.63 0.64 0.68

  31. Slide 31: LSEM – out of sample:

    LSEM — out of sample: Table 2: Forecasting performance — rolling parameter estimates Forecast horizon (in quarters) 1 2 3 #4 5 6 7 8 9 10 11 12 13 14 15 16 Root-mean-squared errors (in percentage points) CPI inflation Naive AR. GDP growth 3-month Libor 0.25 0.40 0.64 0.87 0.94 0.99 0.98 0.95 0.92 0.

  32. Slide 32: DSGE-CH – out of sample:

    DSGE-CH — out of sample: Table 1: Inflation forecast evaluation: Root mean squared errors Horizon 1 2 3 4 5 6 7 8 9 10 11 12 Naive 0.45 0.67 1.03 1.30 148 1.62 1.71 1.78 1.82 1.84 1.83 1.82 DSGE-CH 0.17 0.35 0.52 0.67 0.68 0.64 0.58 0.54 047 041 0.38 0.36

Swiss National Bank — Technical Presentation (May 2011) on 6ix