Assessing the Interest Rate and Bank Lending Channels of ECB Monetary Policies

Transcription

1 Assessing the Interest Rate and Bank Lending Channels of ECB Monetary Policies Jérôme Creel Paul Hubert Mathilde Viennot 1 Paul Hubert, OFCE Sciences Po

2 Motivation (1) Mario Draghi, chairman of the ECB, justified the latest unconventional policy Outright Monetary Transactions, Sept by the disruption of the ECB monetary policy transmission to the real economyinsomeeurozone countries. 2 Paul Hubert, OFCE Sciences Po

3 Motivation (2) This paper aims at establishing the effect of conventional and unconventional ECB monetary policies on both interest rates and lending volumes in the 4 largest economies of the Eurozone during theglobalfinancialcrisis. 3 Paul Hubert, OFCE Sciences Po

4 Literature (1) Most of the literature has focused on: US data the transmission during the financial crisis of ECB monetary policiesthroughthe interest ratechannel only 4 Paul Hubert, OFCE Sciences Po

5 Literature (2) 5 Paul Hubert, OFCE Sciences Po

6 Contribution (1) First, we assess both interest rate and bank lending channels using interest rates andlendingvolumesfor new loans. Second, we investigate the effects of both conventional and unconventional monetary policies; we also disentangle the effects ofonefromthe other. In the spirit of Bachmann & Sims (2012), we artificially shutoff the effects of the QE policy when analyzing the effects of the conventional one, and vice versa. 6 Paul Hubert, OFCE Sciences Po

7 Contribution (2) Third, the analysis is performed, over the financial crisis sample, for the 4 largest economies of the Eurozone: Germany, France, Italy and Spain, andatadisaggregatedlevel encompassing: money markets sovereign bonds at 6 month, 5 year and 10 year horizons loans inferior/superior to 1M to non financial corporations cash and housing loans to households deposits 7 Paul Hubert, OFCE Sciences Po

8 Data Our dataset goes from June 2007 to December 2012 Monthly frequency 67 observations. FromECB StatisticalData Warehouse and National CBsor Treasuries Conventional instrument: ECB rate for main refinancing operations Unconventional instrument: Long Term Refinancing Operations Interest rates and lending volumes for new loans: money markets sovereign bonds at 6 month, 5 year and 10 year horizons loans inferior/superior to 1M to non financial corporations cash and housing loans to households deposits 8 Paul Hubert, OFCE Sciences Po

9 Empirical Method (1) A structural VAR model is used to decompose residuals from the ECB interest rate and QE interventions into mutually orthogonal componentswithastructural economicinterpretation. We augment a standard VAR for monetary policy analysis including industrial production (IP), inflation (CPI) and the ECB interest rate (ECB) with: LTRO operations (LTRO) new loans interest rate (*_r) new loans volumes (*_v) for each market (*) analyzed. We also include in the estimation oil prices,the CISS and STOXX as exogenous variables [they intend to correct estimates for the incidence of the pricepuzzle (oil)andfinancial instability]. 9 Paul Hubert, OFCE Sciences Po

10 Empirical Method (2) Let Z t represent the (k x 1) vector that contains the k endogenous variables atdate t. Inthebenchmark specification: Z t = [IP, CPI, LTRO, ECB, *_v, *_r] Estimated with 3 lags and with a small sample estimator. 10 Paul Hubert, OFCE Sciences Po

11 Empirical Method (3) Our identificationassumptions are: shifts in IP and CPI result in a contemporaneous change in both policyvariables, andprices andvolumesoneach creditmarket. The latter two also react contemporaneously to policy variables, while by construction policy variables would react to innovations to credit prices andvolumesonlywithalag. Concerning the relative position of the two policy variables, we assume that the QE interventions react with a lag to the ECB interest rate consistently with the prevalence of the conventional instrument over unconventional ones. 11 Paul Hubert, OFCE Sciences Po

12 Empirical Method (4) We artificially shut off the effect of one of the two policy variables on prices and volumes of each credit market, in order to isolate the effect oftheother policyvariables oncreditprices andvolumes. 2 conditions required for credit prices and volumes not to react to policyvariables atanyhorizon: Policy variables are ordered before credit prices and volumes in Z t so thatthey donotreactonimpact. We restrict the coefficients on lagged policy variables in equations for credit prices andvolumestozero. These restrictions are implemented by estimating the VAR model using seeminglyunrelated regressions. 12 Paul Hubert, OFCE Sciences Po

13 ECB rate on interest rates Money GB_6m GB_5y GB_10y NFC<1M NFC>1M HH_Cash HH_Hous Deposits Germany (1st row), France (2nd), Italy (3rd) & Spain (4th) 13 Paul Hubert, OFCE Sciences Po

14 ECB rate on interest rates Impulse response in interest rates to a one S.D. innovation increase (a 0.17percentage pointincrease) inecbinterest rate We expect apositiveresponse ofinterest rates tothe ECBinterest rate. The pass through is indeed significant and positive as expected for most countries and markets, except: cash loans to households sovereign bond markets. 14 Paul Hubert, OFCE Sciences Po

15 ECB rate on volumes Money GB_6m GB_5y GB_10y NFC<1M NFC>1M HH_Cash HH_Hous Deposits Germany (1st row), France (2nd), Italy (3rd) & Spain (4th) 15 Paul Hubert, OFCE Sciences Po

16 ECB rate on Volumes We expect credit volumes to be negatively correlated with an increase inthe ECBinterest rate. There is no clear cut pass through over the considered sample from theecbratetovolumes, except for: loans to NFC in Italy and Spain especially those superior to 1M, cash loans to households in Germany and France, housing loans to households in France and Spain, deposits in Germany and Spain. 16 Paul Hubert, OFCE Sciences Po

17 LTRO on interest rates Money GB_6m GB_5y GB_10y NFC<1M NFC>1M HH_Cash HH_Hous Deposits Germany (1st row), France (2nd), Italy (3rd) & Spain (4th) 17 Paul Hubert, OFCE Sciences Po

18 LTRO on interest rates Impulse response in interest rates to a one S.D. innovation increase (a 3.20percentage pointsincrease ofeuro Area GDP) inltro operations. We expect that QE interventions have a negative effect on interest rates (viaincrease inthe moneysupply). No evidence of a pass through from QE to interest rates over our sample, except for : money markets though this is a very small effect, sovereign bonds at the 5 year horizon, loanssuperior to 1M to non financial corporations in Spain. 18 Paul Hubert, OFCE Sciences Po

19 LTRO on volumes Money GB_6m GB_5y GB_10y NFC<1M NFC>1M HH_Cash HH_Hous Deposits Germany (1st row), France (2nd), Italy (3rd) & Spain (4th) 19 Paul Hubert, OFCE Sciences Po

20 LTRO on volumes We expect that QE interventions have a positive effect on credit volumes for the same reason as previously stated: they increase money supply. Once again, there is no evidence of the expected pass through from QE tolending volumes, except for : money markets in Germany, France and Italy cash loans to households in Germany and France. 20 Paul Hubert, OFCE Sciences Po

21 Robustness check: panel data estimations We estimate the equivalent empirical model with panel data: not dynamic, but underlines immediate static effects; univariate rather multivariate, variables are grouped together rather than analysed separately usingtimeseries methods. This enables to control for country fixed effects, potentially important in thetransmissionofmonetarypolicytocredit markets. 21 Paul Hubert, OFCE Sciences Po

22 Robustness: Panel Data 22 Paul Hubert, OFCE Sciences Po

23 Conclusions Main result: Only the pass through from the ECB rate to interest rates has been really effective, consistently with the existing literature, while the transmission mechanism of the ECB rate to volumes and of LTRO operationstointerest ratesandvolumes hasbeen null oruneven. Two argumentstoexplain thedifferentiated pass through: 1. the successful pass through from the ECB rate to interest rates, which materialized as a huge decrease in interest rates, had a negative effect on the supply side of loans, and may have offset its positive effects on lending volumes. 2. capitalheld asexcess reserves attheecbbalancesheet 23 Paul Hubert, OFCE Sciences Po