A large part of the debate on whether or not the ECB can do quantitative easing revolves around the issue that the ECB statutes prohibit the central bank from “financing” any of the member governments directly (or, depending on what German courts say, indirectly as well). To a certain extent this policy makes sense – it avoids a lot of explicit moral hazard and essentially prevents the EU from ever getting stuck in a hyperinflationary situation where governments issue bonds to raise their spending and the central banks just acquiesces and goes on buying these bonds. Also, the Bundesbank has a price stability fetish because of something that happened over 80 years ago but for some reason they can’t seem to learn the correct lessons from. Somewhere else on this blog I have also argued that introducing Eurobonds would provide an instrument for the ECB to actually engage in straight-forward QE, even though just buying a reasonably weighted basket of national bonds would do the same trick (however, with potentially different fiscal implications). But why should buying government bonds be one of the go-to policy to try and gain traction in a liquidity trap in the first place?
Essentially what the ECB needs to do, or in fact what any central bank needs to do in order to avoid the zero lower bound, is to raise the supply of money above the demand for money. Our focus on short-term interest rates, born out of Keynesian, new-Keynesian, but also neoclassical obsession with this supposed indicator of monetary policy (often also mistaken with the price of money), of course makes this hard. The whole liquidity trap argument is based on the notion that at some point “ultra-safe” assets, i.e. government bonds, become perfect substitutes for base money. But there are plenty of other assets out there that have positive interest rates or are not bothered by interest rates at all. McCallum (.pdf) makes the point, also treated later by Svensson (.pdf), that the easiest, indeed a “fool-proof” way of avoiding the zero lower bound on nominal interest rates is to stop looking at nominal interest rates in the first place. This is of course also central to market monetarists thinking on the issue. When nominal interest rates seize to do what they’re supposed to do (which in a neo-wicksellian framework could be described as bringing the real interest rate in the economy in line with the natural interest rate that leads all markets to clear), then throw them out. If the central bank would instead start buying up foreign currency, for instance, thus depreciating the local one and pumping cold-hard cash into the economy, it would be achieving exactly what it wants to achieve (and to a certain degree does with QE): increase the amount of money in the system without having to resort to buying up assets where it strictly speaking makes no real difference whether it buys them or not.
Different to buying government bonds (which, as the liquidity trap argument goes, bear close to 0% interest), it could instead be buying any type of asset with very different nominal interest rates. For all it’s worth the assets bought this way don’t matter as long as they have a higher nominal interest rate (or people who would be willing to hold them for other reasons) than the supposedly ultra-safe asset that is used to claim that monetary policy seizes to be effective once the zero lower bound is reached. Buy gold, buy CDOs, buy other private debt obligations. All of this increases the supply of money in the system and at some point leads to an excess of money supply compared to money demand – and through what is often referred to as the hot-potato effect (people just stop wanting to hold the cash because its quantity quite simply exceeds their willingness to do so) would in turn get the economy rolling. If you want to look at it through the interest rate lens, one might argue that this policy pushes other nominal interest rates down, and given there is clearly a positive demand for those assets (otherwise no market for them would exist and their price would be zero), we can safely assume that in general this lowering of interest rates would in turn reduce the demand for them and push money elsewhere. Sure, there might be some reaching for yield, sure there might be some hoarding, but that’s always the case whenever monetary policy becomes easier, relatively speaking. Just as likely will there be market participants that deem their marginal benefit of consumption to have suddenly become higher compared to the reduced marginal utility of holding the assets that now have lower yields. If the policy works, however, of course the effect would be to raise interest rates in response to a rise in inflation expectations, which is exactly what is supposed to happen. If you think of inflation as being a lowering in the price of money, driving up the prices of assets achieves exactly that – it lowers the price of money in relation to those assets. So if the ECB is barred from buying up government bonds in order to increase the money supply – just let it buy up something else.
EDIT: Removed wrong definition of money supply.