Financial Factors in Business Cycles

1
Financial Factors in Business Cycles

• Lawrence Christiano
• Roberto Motto and
• Massimo Rostagno
• November 23, 2007

2
Model

• Medium scale DSGE model with nominal rigidities
• - Sticky prices and wages
• - Nominal non-state contingent contract
• - Price level reallocate income between
  households and entrepreneurs through Fisher
  debt-deflation channel

3
Model.

• Monetary Policy Adopt a flexible representation
  of Taylor Rule. The percentage deviation of
  variable x is denoted by
• Here, is the actual deviation from steady
  state.
• Target interest rate of the monetary authority
  is
• is set as follows
• Where, is the log deviation of steady state GDP
  growth
• denotes the change in inflation
• is the monetary policy shock, which is
  uncorrelated
• over time.

4
Results

• Estimate model parameters and shocks
• Impulse Response Functions to see propagation
• Variance decomposition to see which shocks are
  important
• Compare models with and without financial
  frictions

5
Parameters governing dynamics

• The structure of their model allows to include
  financial variables such as external finance
  premium and monetary aggregates that are absent
  in PST, LOWW and SW.

6
Estimated Shocks..

• Euro Area
• The vertical distance between actual and
  smoothed data is their estimate of measurement
  error in data. Measurement error is approximately
  zero, showing the similarity between the raw data
  and model predicted data.

7
Estimated Shocks..

• For US The similarity between raw data and model
  predicted data shows the exact decomposition of
  historical data into economic shocks

8
Estimated Shocks in EA..

• Exogenous shock to time t preferences
• below its mean of unity explains the
  increase in savings in EA in the past
  decade
• Banking technology shock at time t
• shock is important in explaining
  growth rate of M1 in EA

9
Estimated Shocks in US

• Banking reserve demand shock at time t
• The sharp spike in 2001 corresponds to huge jump
  in reserves on September 11. This spike
  corresponds to this models spikes in the
  non-borrowed reserves.

10
Dynamic Properties of EA Model.IRF

• Monetary Policy Shock

Nominal rigidity in debt contract gives rise to
Fishers debt-deflation effect. As a result gt
11
Financial Friction Shock

• of exit of entry
• Who exit have more wealth than who enter gt exit
  and entry reduce financial wealth in the hands of
  entrepreneurs as a group.
• As goes up this process is slowed down. With
  additional wealth entrepreneurs buy more capital

12
Marginal Efficiency of Investment

• As increases gt investment is more expensive gt

13
Technology Shock in the Banking Sector

• Technology shock

14
Variance Decomposition to Identify important
shocks
15
Decomposition of GDP Growth for EA

• Dark line indicates actual data, and the bars
  corresponding to each observation indicate
  contribution of shocks.

16
Decomposition of GDP Growth for US
17
Inflation in EA

• Demand shock have the largest impact in EA in the
  first half of the sample.Capital producers and
  entrepreneurs also play a noticeable role.

18
Inflation in US
19
Anatomy of US Boom-Bust

• Simple model is without any banks or financial
  frictions.
• Estimated movements in simple model are exactly
  in the opposite direction from CMR model.
• Simple model over-predicts both Y and I.

20
Policy Implications..

• Investigate alternative monetary policy that
  respond to financial variables such as credit
  growth, stock market and monetary aggregates.
• Standard deviation of output and inflation with
  real stock market growth in EA and US

In EA, central bank (CB) can gain very little by
responding to the stock market. Output and
inflation volatility both rise. In US, CB can
stabilize output and inflation better by
responding to stock market.
21
Policy Implications

• Standard deviation of output and inflation in
  response to broad money growth in EA and US

In EA, broad money is relatively more effective.
In case of US, broad money produces similar
result as stock market.
22
Conclusion

• Interpret the rejection of intertemporal Euler
  equations in the literature in terms of the
  inclusion of the shocks.
• Primiceri, Schaumburg and Tambalotti (PST) (2006)
  will find intertemporal consumption preference
  shock important when they use this model to
  interpret the data.
• Provide a better structural interpretation of
  PSTs results by adding BGG financial frictions.
• Shock emanating from the BGG financial frictions
  is important in understanding the dynamics of the
  data.

23

• Thank You!