Real effects of central bank collateral policy : Liquidity insurance at work

Using a large eligibility shock, I show that liquidity insurance by the central bank has significant real effects. Large-scale rating of non-public French firms, whose credit claims can be deposited as collateral at the central bank, creates a unique opportunity to assess economically relevant effects of collateral policy and identify the transmission channel. The specific structure of available data makes it possible to tackle the usual shortcomings of Differences-in-Differences estimators and the problems identified in the recent econometric literature. As regards the effects, eligibility extension significantly increased material investment and employment, especially for small firms more affected by liquidity constraints. As regards the channel, firms were able to invest more because they received more long-term loans; an arbitrage argument shows that banks were able to lend long since they enjoyed liquidity insurance by the central bank. Although a large literature analyses bank runs and the effects of liquidity on long-term lending and real outcomes, this seems to be the first study providing microeconometric evidence on the direct financial and real effects of liquidity insurance à la Diamond and Dybvig (1983). Finally, thanks to the multiple layers protecting its balance sheet, the central bank did not take over firms’ default risk; this feature makes it possible to discard a “risk-inflation” channel. These results, which diverge from part of the literature, depend crucially upon a rigorous institutional setting and a clear distinction between collateral refinancing and direct assets purchases by the central bank (Quantitative Easing)