In this second part of my take on liquidity traps and missing collateral in global financial markets I would like to respond to some of the talking points set out by FT Alphaville's Cardiff Garcia (again in response to the much talked about piece by Credit Suisse). Even though blog posts tend to age quickly, recent central bank action suggests that this topic is relevant as ever. Specifically, negative readings across a wide range of short term interest rates in Europe has raised the question not only what the effect of such abnormal interest rates are, but also whether such market prices are sending a signal to central banks that they ought to act much more aggressively.
Following up on FT Alphaville's coverage, one question that is intereting to consider is the following.
2) The movement of M1 and M2 in recent years seems not to have told us anything helpful about inflationary prospects. Should the Fed finally ditch them and start concentrating on another measure, perhaps one that incorporates some of the items above? Or bring back M3 (which at least included such shadow banking elements as institutional money market funds and repo)?
Initially, I should point out that I disagree with the premise of this statement. I think that there are still important effects from fluctuations in M1 and M2. Specifically, I believe that while being in structural process of deleveraging may certainly mitigate the inflationary pressures from central banks generating excess reserves (and liquidity) in the system, it is dangerous to assume that expanding base money does not have a real economic effect.
Still, this raises a very important point. The traditional monetary policy transmission mechanism is broken and as a result the size and expansion of base money aggregates have little bearing on credit creation in the real economy. The key question is the whether central banks should extend their control of the money supply further down the credit foodchain (i.e. closer to the end user/beneficiary of the credit)? And if you answer to this is yes, how do central banks do this most effectively.
Firstly, it is important to emphasize that, in many ways, they already have. Initial responses to the crisis in the US (and QE conducted by the BOJ) have been engaged in strategic and direct purchases of several kinds of marketable debt and equity securities, but central banks generally do not like to do that. Historically, the Bank of Japan has been most direct trying to influence market prices through the purchase of corporate bonds and exchange traded funds.
Now however, they are at it again of course. The BOE recently suggested open market operations with strings attached in the form of banks only getting access if they added to their balance sheet and in Europe, the ECB has cut its deposit rate to 0% and may even push it into negative this week.
The problem however is that it is very complicated for the central bank to do this effectively and a central bank will always be adverse to taking direct market risk (even if e.g. the allegedly most conservative central banks of them all, the ECB, has taken substantial market risks through the collateralised LTROs). In addition, targeting M3, M4 etc would mean an even more direct involvement in the credit process by which the central bank potentially acted as direct underwriter for pools of securitised loans of all shapes and sizes. This adds illiquidity to the balance sheet and exposes the central bank to significant mark to market risks which eventually may have to be covered by printing money. Such an implicit backstop to securities that the central bank may agree to buy creates significant moral hazard.
Traditionally, a central bank will respond to a liquidity trap by supplying (potentially) unlimited levels of excess reserves to the banking system and thereby expanding the potential credit supply in the economy. The counterbalancing asset side entry here will usually be short term government bonds but also, if need be, longer term government securities. In this sense, expanding the balance sheet at the zero bound is essentially a fiscal expansion. However, as Izabella suggests, this may actually be counterproductive in an economy suffering from a structural lack of liquid and investment grade collateral.
The central bank will then actually exacerbate the lack of such assets by doing textbook QE which involves creating bank reserves in exchange for short term government securities. Demand for government securities (collateral), the story goes, would be more than enough to keep yields down and allow the government to conduct fiscal expansion at the zero bound. Still though, one would have to assume a complete lack of any market response from bond vigilantes ad infinity for this to work. I am not sure that I accept this.
So where does a broader monetary aggregate target come in? Well, from the account above the central bank could do two things.
1. Act on the liability side by aggressively cutting excess reserve requirement and enforcing a penalising rate on excess reserves. This would be a direct way (through the liability side) to attempt to jump start the money multiplier and force up velocity, but it will require the central bank to be indifferent between currency and reserves on its liability side (see below).
2. Avoid crowding out demand for safe collateral by booking anything but government securities on the asset side.
On the second point, I would note that this assumes a complete lack of bond vigilantes of any kind and thus no disciplinary market action in the context of financial repression. In the context of structurally overlevered governments across the developed world, I am not sure that this is a reasonable assumption over time. What would Gilts be trading at if the BOE was not holding 30% of the total stock outstanding and how would this have affected the government's ability to borrow. In addition, taking direct market risk by e.g. purchasing pre-assigned tranches of securitised loans (to beef up broader monetary aggregates) would certainly not work if the underlying problem was one of a structural lack of solvent credit demand. I have argued for example that this is a major part of the problem both in the context of private and public borrowers.
On the first point, events have caught up with theorizing here with the ECB the first major central bank now imposing zero interest rates on its deposit rate (the Riksbank did this in 2008/09 too) and there is a serious probability that the rate may be moved into negative.
I have been lucky to have the opportunity to discuss this with the financial columnist and investor Sean Corrigan who makes the following crucial point.
People are treating this [negative deposit rates] in a completely erroneous fashion. Even negative rates cannot force the banks to lend out the deposits they hold at the central bank; this is to assume they - as a whole - have choice in the matter: they do not. If the central bank creates what is called 'outside' money, by buying securities, etc, the corresponding reserves cannot be voluntarily removed: they have to be held as CB liabs/commercial bank assets on the central bank balance sheet.
What such a move could possibly do is to make it more imperative for banks to leverage up by creating new, extra loans (To whom? On what terms?) and to accept new customer depo's (given that they hold a huge surfeit of reserves on their balance sheets with which to backstop these) and so compensate for the negative rate drain by the volume effect. I take this as dubious, however, given normal credit concerns and, in some cases, binding capital constraints.
Sean then goes on to make the final point that if the central imposes negative interest rate on reserves while at the same time wanting to maintain the size of its balance sheet, it would have to generate currency as reserves were withdrawn.
I think a couple of points are important to note at the offset.
The situation at the ECB and the Fed/BOE is different. More specifically, reserves created in the Fed/BOE is, as Sean points out, "outside money" and is thus what we could call "push QE". The central bank has a policy objective to affect government bond yields and the only way it can do this is to generate reserves in the system. At the ECB and while the central bank may certainly have intended to affect government bond yields in the periphery this was "pull QE"; i.e. reserves were pulled from the ECB based on banks' demand for such liquidity and, presumably, their need to shed themselves for collateral.
One of the main objectives as stated by the ECB was to provide liquidity to banks who could not otherwise refinance themselves in the short term money market. I think it is critical here to note that while the Fed and the BOE have always put forward specific targets for their QE operations, the ECB has not! If the banks had put up 2 trillion worth of collateral in the LTRO and asked for 2 trillion in liquidity they would have gotten it!
Anyway, to the point that banks are forced to hold these reserves (as a counterpart to the size of the balance sheet), in the perfect world it is not certain that this is the case. Let us assume a central bank conducts QE through the expansion of reserves with two objectives in mind.
1) To affect government bond yields (perhaps to allow the sovereign to run higher cyclical deficits for a period to boost aggregate demand).
2) To improve risk sentiment and risk taking in the economy through higher credit creation by commercial banks.
One immediate effect of such a policy in an environment where the monetary policy transmission is broken is that while government bond yields may go to zero it has no effect on lending to the real economy.
What can the central bank do? Not a whole lot as it were.
The central bank created the reserves in the first place and cannot easily induce banks to lend these out without compromising its asset side. In other words, it is difficult to pursue both objectives directly at the same time.
Specifically, if banks started to draw on its excess reserves to lend out to the real economy the central bank would need to one of two things. Maintain the size of its balance sheet constant by creating currency or reducing its holdings of securities on the asset side (government bonds). The latter is difficult to do especially if the central bank has been very aggressive in its sovereing bond purchases. Still, theoretically the central bank will be informed by the notion that the economic multiplier of commercial banks lending out to the real economy is higher than central bank financed government spending. It is not inconceivable that this is what central banks may start trying to do. We are seeing this by the BOE now giving preferential treatment to banks lending out and the ECB with its latest move.
Finally, negative rates could, as noted above, force banks to lever up to make money and it is not inconceivable that this could work in some countries where the banking system sounder or where some banks may have the buffer to do so. The main risk for the central bank is that this reduction in its balance sheet simply forces reserves into government bonds anyway as commercial banks see no other choice. The structural features of financial repression also will point towards this.
Here of course, the practicality becomes an issue. The BOE now holds 30% of all Gilts outstanding and if it started to sell these off as banks started to lend to the real economy interest rates across all maturities and lending products could rise very fast and in complete disconnection with underlying fundamentals. Currently, the position for central banks in this matter is complicated because as we have seen liabilities tend to migrate to the government balance sheet and as such the central bank needs to work very hard to keep borrowing costs in check for the sovereign.
Now, as for creating physical currency it is not clear to me that central banks would want to do this but the notes that I have read so far simply assume that it is given that commercial banks would be able to shift reserves into currency. This then brings up a whole hosts of other issues regarding the existence of cash at the zero lower bound. Citi's chief economist Wilhelm Buiter for example has suggested that physical currency be retired altogether and that electronic money be used instead. Such electronic money could of course be subject to exactly the level of negative interest rates the central bank deemed fit or perhaps even be subject to a fixed maturity.
Negative deposit rates have another effect in so far as they induce carry trades in with the negative currency yielder as funder and thus pushes the currency (euro) down. Such real effective depreciation could be a powerful tool for the ECB and for once it is a first mover here.
Ultimately, negative deposit rates are no panacea and certainly in the context of central bank creating the excess reserves in the first place, but the effect of FX markets as well as the potential for its effect on the quantity of currency in the system should be keenly watched.