The financial crisis that started in 2007 has seen central banks play an unprecedented role both to ensure financial stability and to support economic activity. While the importance of the central bank in ensuring financial stability is well known (see e.g. Padoa-Schioppa (2014)), the unprecedented nature of the financial crisis led central banks to resort to new instruments for which the literature offered little guidance. This thesis aims to bridge this gap, using both theory and data to better understand one of the main instruments used by central banks: collateralized loans. The general contribution of the thesis is thus both retrospective and forward looking. On a retrospective point of view, it helps understanding the actions of the central bank during the crisis and the mechanisms involved. Looking forward, a better understanding of the tools used during the crisis allows to better inform future policies.The first chapter starts from the observation that the literature, starting with Bagehot (1873), has generally assumed that the central bank should lend against high quality collateral. However in the 2007-2013 crisis central banks lent mostly against low quality collateral. In this chapter, we explore when it is efficient for the central bank to relax its collateral policy. In our model, a commercial bank funds projects in the real economy by borrowing against collateral from the interbank market or the central bank. While collateral prevents the bank from shirking (in the spirit of Holmstrom and Tirole (2011)), it is costly to use as its value is lower for investors and the central bank than for the bank. We find that when the bank has high levels of available collateral, it borrows in the interbank market against low collateral requirements so that the collateral policy of the central bank has no impact on banks' borrowing. However, when the amount of available collateral falls below a threshold, the lack of collateral prevents borrowing. In this case, the collateral policy of the central bank can affect lending, and it can therefore be optimal for the central bank to relax its collateral requirements to avoid the credit crunch.The second chapter focuses on collateralized loans in the context of the euro area. According to the literature on optimum currency area, one of the main drawbacks of currency unions is the inability for the central bank to accommodate asymmetric shocks with its interest rate policy. Suppose that there are 2 countries in an economy and one suffers a negative shock while the other has a positive shock. Theory would suggest an accommodative policy - low interest rates - in the first country and a restrictive policy - high interest rates - in the second one. This is however impossible in a currency union because the interest rate must be the same for both countries (Mundell 1961, McKinnon 1963, de Grauwe 2012). In this chapter I show that collateral policy can accommodate asymmetric shocks. I extend the model of collateralized lending of the first chapter to two banks A and B and two collateral types 1 and 2 .I also introduce a central bank deposit facility which allows the interest rate instrument to be compared with the collateral policy instrument in the context of a currency area hit by asymmetric shocks. Macroeconomic shocks impact the investment opportunities available to banks and the value of their collateral and the central bank seeks to steer economy rates towards a target level. I show that when banks have different collateral portfolios (as in a monetary union where banks invest in the local economy), an asymmetric shock on the quality and value of their collateral can increase interest rates in the country hit by the negative shock while keeping them unchanged in the country with a positive shock.The third chapter provides an empirical illustration of this “collateral channel” of open market operations. We use data on assets pledged by banks to the ECB from 2009 to 2011 to quantify the “collateral substitution / smoother transmission of monetary policy” trade-off faced by the central bank. We build an empirical model of collateral choice that is similar in spirit to the model on institutional demand for financial assets of Koijen (2014). We show how the haircut of the central bank can affect the relative cost of pledging collateral to the central bank and how this cost can be estimated using the amount of assets pledged by banks. Our model allows to perform a broad set of policy counterfactuals. For example, we use the recovered coefficient to assess how a 5% haircut increase on all collateral belonging to a specific asset class (e.g. government bonds or ABS) would affect the type of collateral used at the central bank. The final chapter focuses on the use of loans as collateral by banks in the euro area. While collateral is generally viewed as consisting of liquid and safe assets such as government bonds, we show that banks in Europe do use bank loans as collateral. We identify two purposes of bank loan collateral: funding and liquidity purposes. The main distinction between the two purposes is with respect to the maturity of the instruments involved: liquidity purposes refer to the use of bank loans as collateral to obtain short term liquidity and manage unexpected liquidity shocks. In practice the central bank is the main acceptor of these collateral. The second type of use is for funding purposes, in which case bank loans are used as collateral in ABSs or covered bonds. The collateral in these transactions allow banks to obtain a lower long-term funding cost. / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished