Transcript Slide 1

The Zero Lower Bound, ECB Interest
Rate Policy and the Financial Crisis
Stefan Gerlach and John Lewis
Discussion
Gert Peersman
Ghent University
Overview
• Estimation of a monetary policy reaction function for the ECB
– Over the period 1999-2009
– Lagged interest rate, lagged change in interest rate, PMI, HICP inflation,
M3 growth, nominal effective exchange rate
– Allow for a smooth transition from one set of parameters to another
• Transition based on time trend or output growth
• Findings
– A break in the months following the collapse of Lehmann Brothers
– Most variables (except M3 growth) significant in pre-crisis period, only
lagged interest rate significant anymore in post-crisis sample
• Interpretation of different interest rate setting
– ECB cuts interest rates more aggressively since crisis, in line with
optimal monetary policy in the vicinity of the zero bound
Discussion
• Very interesting and relevant topic
• Well executed and clearly written
• My comments/remarks
– Interpretation of results
– Methodology
Interpretation
• Gerlach and Lewis: ECB cuts interest rates more aggressively
since crisis, in line with optimal monetary policy in the vicinity
of the zero bound
– Because actual interest rate since
Lehmann collapse lower than
predicted interest rate based on
pre-crisis reaction function
• Does this finding really proves that the ECB reacts more
aggressively when reaching the ZLB?
– Notice: none of the parameters are significant in post-crisis reaction
function (except lagged interest rate)
– Shouldn’t we rather find more responsive parameters when the
likelihood of reaching the ZLB increases?
Interpretation
• Alternative interpretation of results: omitted variables problem
– Central bank has much more information at time of policy decision than
two month’s lag of inflation, two month’s lag of M3 growth, one month’s
lag of PMI and one month’s lag of the exchange rate
• There was a crisis on financial markets and collapse of Lehmann itself: ECB
knew that this shock was going to be transmitted to the real economy, and
reacted accordingly
– Policymakers looking at several other variables (spreads, asset prices,
bank liquidity) rather than conventional output and inflation indicators
• Output and inflation forecasts not very reliable during that period and often
revised: could be explanation why none of the parameters of conventional
variables are significant anymore in post-crisis reaction function
• Supported by earlier literature (e.g. Rudebusch 2001, Orphanides 2003,
Aoki 2003): increased uncertainty about an information variable implies a
more cautious optimal response to that variable
– Peersman and Smets (1999): optimal coefficient in reaction function
can even be zero when uncertainty is very high
Interpretation
• Alternative interpretation of results: omitted variables problem
– Careful: high estimated degree of interest rate smoothing for pre-crisis
period (0.98) is key to have a higher interest rate prediction for post
Lehmann period
• OK if persistence of policy rates reflects
deliberate inertia of the central bank, i.e.
intrinsic inertia
• Rudebusch (2006): estimated persistence
could reflect the response to slow cyclical
fluctuations in key macroeconomic driving
variables of monetary policy or omission of
important persistent influences on actual
setting of policy, i.e. extrinsic persistence of the policy rate
• Difference is important: “For example, when faced with a surprising
economic recession or a jump in inflation, the inertial policymaker slowly
changes the policy rate, while the non-inertial policymaker responds to the
news with immediate and sizeable interest rate adjustments”
Methodology
• Monetary policy regime (switch) is a deterministic function of
time, and only one transition is possible
• Why not using a Markov-Switching Model with parameters that
could change across regimes?
– More general approach, nesting more possibilities
– Monetary policy regime endogenously determined
– Data determines whether there is only one regime-switch in the sample:
transition probabilities across regimes are estimated and could even be
one or zero, i.e. there is no need to switch back to earlier regime
– Possibility to have more regimes
– E.g. Sims and Zha (AER, 2006)
Methodology
• Extension: transition between regimes depends on GDP growth
– Economic variable determining the switch, which could also be reversed
• But why GDP growth to affect ECB reaction function?
– P 17: “These reasons suggest that the PMI is likely to be more strongly
correlated with the ECB’s view of real economic activity than real GDP,
and therefore more suitable for inclusion in the reaction function”
– If you want to examine different monetary policy behavior when interest
rate approaches ZLB, why not allowing regime to be dependent on
level of the interest rate?
• Note: more aggressive policy response
already at interest rate level above 3%
• Interest rate has actually been closer to
ZLB before (e.g. 2% in 2003-2005)
Methodology
• Monetary policy stance involves more than repo interest rate
– Peersman (2010): ECB has more than one instrument that could affect
the real economy
• A 10 basis points innovation in repo rate has a similar impact on output after
one year as a 2 percent exogenous decline of the monetary base that is
orthogonal to repo rate and shocks to demand for monetary base
Output
Monetary base
ECB minimum bid rate
0.12
0.4
0.50
0.08
0.0
0.25
0.04
-0.4
0.00
-0.8
-0.04
-1.2
-0.08
-1.6
-0.12
-2.0
-0.75
-0.16
-2.4
-1.00
0.00
-0.25
0
12
24
36
-0.50
0
12
24
36
0
12
24
36
Methodology
• Monetary policy stance involves more than repo interest rate
– Peersman (2010): ECB has more than one instrument that could affect
the real economy
• A so-called Monetary Conditions Index could be calculated
5
5
4
4
3
3
2
2
MCI
1
1
MP rate
0
0
-1
-1
1999M01
2001M01
2003M01
2005M01
2007M01
2009M01