QE, Monetary Financing, Bank Money and Anarchy
by George Hatjoullis
The thrust of recent blogs is that the global economy is in a ‘Great Deflation’ despite pronouncements to the contrary by central bankers. Others refer to secular stagnation but I believe the latter term best describes a different structural problem. The ‘Great Deflation’ captures the monetary nature of the present policy dilemma and its primary roots in the banking crisis of 2008. One of the disadvantages of this formulation is that everyone looks to aggregated price indices for evidence of the condition. Aggregated price changes are merely symptoms of the underlying disease and as with physical illness, presenting symptoms do not always signify the seriousness of the condition. Central banks, notably the Federal Reserve, focus on surveys of what people think will happen to aggregated prices. They used to also look at inflation expectations implied in bond yields. However, these were delegated to ‘compensation for inflation risk’. The market is requiring less compensation for inflation risk than the Federal Reserve believes it ought to be. This is the thrust of Janet Yellen’s recent speech and covered in detail in a recent blog post. This is a very odd situation. A central bank telling the market it is wrong and that it ought be requiring more compensation for inflation risk. The market, however, seems unmoved and confident in its view. One does not have to be a die-hard advocate of the Efficient Markets Hypothesis to suggest the Fed is ignoring some important information here. There is a reason the market is not getting anxious about inflation, despite the protestation of the Fed. Explaining the position of the market has been the purpose of my repeated blog posts on the ‘Great Deflation’. The fact that the market is unmoved is itself evidence of a deflationary bias in the global system at the moment. The purpose of this blog is to revisit (yet again) this discussion in the context of the Iceland proposal for Sovereign Money.
First, a quick note on inflation indices. Various blog posts have discussed aggregated price indices and their limitations. One observation has been that the only meaningful aggregated price index is the GDP price deflator. This has the added advantage that it is broadly comparable across nation states. Consumer price Indices are not comparable in all cases. If the US CPI and so-called Core CPI were recalculated using the same method as the EU HICP series it is quite probable that it will reveal that US inflation is lower than it seems. Markets tend to overcome such opacity with remarkable clarity and this may explain much.
The story so far is that central banks resorted to quantitative easing because the conventional mechanism for expanding the money supply, bank money creation, was ineffective. The banks were flooded with plentiful, cheap, reserves but failed to expand credit, and thus bank money, sufficiently. Inflation rates continued to drift lower. The central banks attempted to circumvent this by purchasing sovereign debt (and other assets) directly in the market. In principle, this replaced risk free assets (government liabilities) with another risk free asset (central bank liabilities). However, since the source of the problem was the credit risk aversion from the fallout of the 2008 crisis, the action had little direct benefit. In fact it may have made matters worse. The new-found recognition that uninsured bank deposits were not risk free created a dilemma as to what to do with the central bank liabilities received in exchange for the risk free government debt. Once deposited with a commercial bank the sums received from the central bank in exchange for debt ceased to be risk free over and above the insured amount (£85k in the UK). QE reduced the quantity of nominally risk free assets ay a time when there was excess demand for risk free assets.
The central banks assumed that simply by removing risk free assets from the system they would encourage economic agents to willingly hold riskier assets. This was evidently not the case. It seems that economic agents chose to buy what risky assets were left at higher prices. Much of the outstanding stock of risk free assets was held by commercial banks and, to the extent that economic agents bought the government debt at inflated prices from the commercial banks, the central bank liabilities were returned to the central bank. Commercial bank government bond sales simply made the already liquid banks more liquid and excess commercial bank reserves are held with the central bank. The size of this QE ‘leakage’ is unclear but it must have existed. The remaining central bank liabilities injected via QE by definition ended up in riskier assets. The net effect of the greater purchases of risky assets, and lower risk free interest rates, was to boost asset prices, notably property and equity. It was through this ‘wealth effect’ that QE had its positive economic impact.
Nominal interest rates and bond yields remain low to negative. The path of future interest rates implied by the term structure is too low according to central banks. In other words and as noted above, the market required compensation for inflation risk is too low. This is an odd situation. The central banks are telling the market to be worried about inflation and the market is unmoved. The reason is that the market does not have sufficient credit risk free assets, in part because the central banks have removed the assets and in part because of the awareness that uninsured bank deposits are no longer to be viewed as wholly risk free. Inflation risk, it seems, is judged by the market to be the lesser of the two evils. Central banks, notably the Federal Reserve, seem set to raise interest rates at a faster pace than the market believes can be sustained. One of the two will ultimately be proven wrong.
If the market is proven wrong then a bond bear market of some ferocity is about to be unleashed. If the Federal Reserve tightens too quickly and by too much, it will reinforce the evident deflationary bias. Both outcomes will be economically disruptive but from a financial stability perspective it is best if the market is wrong and not the Federal Reserve. One of the interesting aspects of the Iceland Sovereign Money proposal is that this awkward situation cannot arise. This is worth exploring.
Let us introduce a Transaction Account (TA). This account is operated by banks but represents electronic cash. The cash is a liability of the central bank and not of the operating bank. It may charge a fee but otherwise it operates like a current account. Funds deposited in a TA are not available to finance bank lending. There is no need for deposit insurance because the Federal Reserve stands behind the account. Typically they will pay no interest and fees will imply a negative interest rate. However, there is no limit on how much can be held in such accounts and they are risk free. If the operating bank fails, your money is safe. The banks will also offer Risk Accounts (RA). These are uninsured deposits with the bank. They are a loan to the bank. If the bank fails your money in these deposits is at risk. If, as I have asserted, there is a shortage of risk free assets then money will flow out of RA into TA. Banks will start to offer higher rates on these Risk Accounts to attract deposits unless they can get cheaper funding from the Federal Reserve. Let us now assume that the Federal Reserve introduces a new rule that restricts the amount of bank assets that can be funded by central bank reserves. It does not need to break the link altogether as in the Iceland proposal but merely restrict the available liquidity under all circumstances. Commercial banks now have to reconsider their funding plans.
At the moment banks can get funding from the Federal Reserve at a price so it is a matter of managing costs. If however there is a physical limit on how much funding they can have, irrespective of price, it becomes a different optimisation problem. They need to be sure they can fund. This will force them to cultivate a Risk Account deposit base and this will involve paying consistently attractive interest rates on such deposits. There is no necessary systemic risk implication because the TA allow concerned depositors to keep funds safe if they choose. RA are explicitly investments and not bank money. In this arrangement the rate at which banks can borrow from the central bank becomes less important because they need to fund themselves in the market as well. The rate of interest becomes market determined. If the banks want to make a lot of risky loans they have to entice money out of TA into RA and RA deposit rates rise. In terms of Janet Yellen’s speech the equilibrium real rate of interest becomes market determined and does not depend on the Fed having a crystal ball.
The interesting question is how then does the Fed or any CB influence the money supply, (i.e. the stock of TA deposits) in this structure? First, note QE now becomes more effective. If the CB buys government debt it places CB liabilities directly in the system and there is somewhere for the sellers of the debt to place the proceeds that is risk free, namely the TA. If all the proceeds go directly into TA then the money supply increases by the amount of the QE. If some of the proceeds go into RA deposits then the banks can expand credit (which would presumably be the point of QE). Eventually, the funds would find their way back to someones TA so the money supply expansion is always equal to the amount of the QE. It cannot be more because bank deposits (RA) are no longer money. A moments thought will reveal that this seriously limits the potential for monetary growth. How does the CB expand money supply faster if it deems it necessary?
Limited monetary financing is the solution according to the Iceland Sovereign Money proposal. The CB sets a nominal GDP target based on some inflation target in terms of the GDP deflator and potential for GDP growth. It calculates the amount of money it deems it necessary to achieve this target on average, over time and gives it to the government of the day to spend. The money supply increases by this amount. It is not open to abuse because the CB is independent of the government of the day and has a specific target. The benefits of money creation, seigniorage, accrue to the state and not the banks.
The flaw in this set up is the implicit assumption that CB money is the only money. Let us assume I become a big electronic money operator. Let is call it eCoin. People keep eCoin with my exchange and transact through my exchange. I have a large stock of eCoin and everyone starts to trust my exchange. I start lending eCoin secure in the knowledge that I can cover any daily net flows out of my exchange. As long as my exchange can cover net daily transfers to other exchanges, my loans can expand to well beyond the amount of eCoin that I own. If the exchanges combine to clear net proceeds at the end of the day and all transactions are covered through exchanges then eCoin can become a private money with the capacity to grow, through leverage, beyond the coin produced. Such a parallel money can grow to dominate the CB money and fund transactions, and inflation, beyond targets.
The flaw in all market based systems is the assumption that they can somehow be controlled. Markets are the ultimate in anarchy; order without a central control. If the economic system wants more lubrication than the official lubricators will provide then, like nature, it will find a way. There is a lot of merit in the Iceland proposal but it is not a panacea. The problem lies in the belief that a system exists that is universally valid over time and context. Monetary systems need to be tested regularly to see if they are still fit for purpose. Dogmatism tends to rule such systems rather than pragmatism and that is why we get so many crises. Anarchy rules in the market place and it can never be controlled only diverted for a while.