An estimated Euro area model with financial intermediation and central bank liquidity operations Discussion by Massimiliano Pisani Bank of Italy CBMMW 2013 Central Bank of the Republic of Turkey Istanbul, 7-8 November 2013 USUAL DISCLAIMERS HOLD
What’s the paper about • Monetary explanation of the business cycle • To which extent liquidity management by the banking sector matter for business cycle? • Estimation of an (otherwise standard) new Keynesian model featuring a non-trivial role for liquidity
The transmission mechanism • Banks finance loans to firms with households’ deposits and liquidity reserves provided by the central bank • Banks need eligible: public sector bond is collateral in open market operations for getting reserves • Stylized banking costs (increasing in the amount of loans/decreasing in the amount of reserves)
The transmission mechanism • The interest on loans depends on • monetary policy rate • banking costs • value of public sector bonds, which are the collateral in the liquidity constraint of the banking sector
The transmission mechanism: implications • Pass-through from the policy rate into the rate on loans is not full (rate on loans different from Taylor rule based policy rate) • Banking costs do matter for rate on loans • Public debt provides liquidity services (Ricardian equivalence does not hold, because public debt is a collateral and hence distortionary) • Central Bank can affect real allocations not only through the policy rate (Taylor rule), but also through banking sector reserves
Comment: banks’ collateral constraint • Key equation: amount of liquidity the banking sector gets from the central bank depends on the value of collateral, the government bond It = ktpBtBt-1/Rmt • Above mechanism is theoretically parsimonious and powerful (public debt becomes a relevant variable for liquidity, the constraint amplifies shocks)
Comment: banks’ collateral constraint • For the estimation, a shock ηomo,t is included: It =ηomo,t ktpBtBt-1/Rmt • According to the results, ηomo,t is a relevant determinant of reserves and lending rate (forecast error variance decomposition)
Comment: banks’ collateral constraint • Comment: shock in the constraint should be further discussed • Motivation for ηomo,t: according to the authors, in the data only 1/3 of collateral is government bonds; is this a relevant source of misspecification? • Moreover, ηomo,t could also capture the constraint slackness (is the constraint always binding in the data? • No easy theoretical solution (adding bonds other than those issued by public sector in the constraint is not trivial) • Comment: Why k is set to one (money injections vary with price of collateral)? Is it consistent with policy practices in the estimation period?
Comment: Cost of financial intermediation • Results suggest shocks to cost of financial intermediation is relevant for deposit and lending rates Other existing (micro) evidence on its properties: • relevant determinant of the pass-through of policy rate into interest rates on loans and deposits? • Functional form and calibration?
Comment: public finances • Public sector bonds provide liquidity services • Interaction between public finance and monetary policy rules (on money injections and policy rate) can affect (1) price stability, (2) equilibrium determinacy and, hence, (3) estimation • More discussion is needed
Comment: variance decomposition results • Looking at charts and tables, there is a “disconnect” between liquidity sector and real variables • New banking sector-specific shocks do not greatly contribute to explain the variance of real variables • New banking sector-specific shocks explain mainly banking sector variables • Is this result model-specific? Or is it common to other models featuring financial imperfections?
Suggestion: Stylized facts, moment matching, sensitivity • Report stylized facts/puzzles of the market for liquidity (loans, deposits,..) • Report moment analysis (variance, persistence, cross-correlations) • Report sensitivity analysis
To conclude • The paper is an attempt to take interesting theoretical features of monetary policy business cycle to the data • I enjoyed thinking about this paper
To conclude • While the theoretical model seems fine, authors should better think about the estimation/quantitative side (the current draft is preliminary, of course) • Quantitative results built on (1) sound calibration, (2) analysis of the interaction of monetary/public finance policies would be equally interesting • Looking forward for the follow up