Monthly Archives: August 2016

Inflation, QE, money, banking and other stuff

Inflation might be always and everywhere a monetary phenomenon, but the devil is in the detail. Inflation as such is a social convention, which usually means the annual change in the consumer price index (CPI). One might as well speak about nominal GDP growth or some other nominal measures (NGDP would be a far better target for CBs, as future expectations affecting investment are tied to future profit expectations). Since 1990’s, the central banks have been more or less independent and the usual wisdom is to keep inflation stable, around 2%. The independence I guess is basically there to prevent excess monetary financing of government deficits. And I stress the word ‘excess’.


The recent ‘unconventional’ monetary policies in the US, UK, Japan and Eurozone have included measures such as buying ABSs, government, covered and corporate bonds. The interest rate corridor is ultra-low, with interbank rates around zero or even negative. Eurosystem has utilised a negative rate on central bank deposits (-0,4 % interest on banks’ checking accounts at the central bank). So the price of money is low, measured by the yield curve. The yield curve is therefore rather flat as well.


When QE was launched by the Fed  in the US some years ago, some investors and economists claimed that QE would lead to  hyperinflation or at least would undermine the external value of dollar to a great extent. Both are theoretical possibilities, but I think that the common and usual discussion around money, inflation and monetary policy is somewhat shallow and unrigorous. For sure,  inflation and monetary economics are difficult subjects and much studied in the economics literature (especially in the form of modern DSGE models). However, I think that one of the main reasons for the misconceptions around inflation is the flawed ontology around money creation and purchasing power.


The usual claim is that central bank bond buying is somehow affecting inflation directly (CB money printing causes hyperinflation, therefore buy gold or stocks). There is an effect, but it is mostly indirect. Let me sketch a simple model economy.


We assume there is a central bank, a private bank (the banking sector is consolidated into one bank), some firms and households. We also assume there is a government with some stock of public debt D. Let us assume that the interest rate level is already at the so-called zero lower bound. Now, as the economy is in secular stagnation, the central bank records a zero inflation development from the statistics office and decides to launch an ambitious QE programme. We assume that the private bank holds all the government bonds of amount D. The private bank has a checking account at the central bank with a balance of 0, for simplicity. The central bank then buys the whole bond stock D, worth of 100 from the bank and credits the bank’s central bank account by 100. This is QE. The direct effect is that the balance sheet of the CB expands by 100 and that the return on equity for the bank goes down (checking account at the CB yields less than the government bond (hopefully ;)).


Hold on, where’s the (hyper)inflation? Well, as consumer price inflation depends basically on the nominal purchasing power of households, the additional inflation requires additional nominal purchasing power for the households. For sure, the balance sheet of the central bank has increased by amount of 100 and maybe the yield curve has flattened a bit (portfolio effect, as the bank wants to swap the low yielding CB deposit into some other fixed income instrument). This is the more or less direct effect of QE. But the fact that the private bank holds now 100 units of central bank money instead of government bonds does not add to the purchasing power of households.There is the small effect of monetising the government interest costs (government owns the CB), but this is of minor importance.


The purchasing power of households will increase, if they borrow money from  banks (consumer credit) or if their salaries increase (or if they receive dividends/sell assets). Now, let us assume the ‘salaries increase’ -scenario. This can take place through investment (hiring more people in the aggregate) and/or additional consumption in the economy (more revenue for the firm means a possibility for negotiating higher wages). This in turn requires that more purchasing power is injected in the economy.


This is where banks enter the equation. Let’s assume that a firm wants to invest by amount of 100 and borrows the money from a bank. Then the bank credits the firms checking account by 100 and recognises a new loan on its loan book. Nothing else is required. The fact that the bank had 100 units of central bank money on its balance sheet instead of government bonds, due to QE, did not affect the loan agreement in anyway directly. This is due to the fact that in a closed economy, banks do not lend their excess reserves (that’s the reason btw why there is no hyperinflation in the US in spite of the huge pile of excess reserves). The excess reserves can in this case only go down by extending new loans, creating therefore new deposits and forcing down the reserve-ratio (CB checking account balance/banks checking accounts balance). The same logic goes for consumer credit or mortgage loans, new deposits are created and therefore new money is created. Out of thin air, not from the CB deposits.


Therefore demand for loans is critical. Of course low interest rates help, but if the economy is excessively leveraged, the deleveraging dominates and there is no demand for new loans in net terms (too much private and public debt). Also, banks are constrained by the amount of regulatory capital, so little capital=little new lending.


Now, as money is commonly understood to be cash and banks’ checking accounts, it is quite clear that new money in this respect requires new lending from the banking system. QE does not increase directly the balance sheet of the banking system. It only flattens the yield curve and creates excess reserves (and makes the government debt interest -free as the government receives the interest income from the central bank in the form of dividends).


To sum up, QE should create strong incentives for new lending and borrowing, especially now, when excess reserves are penalised. Despite of low interest rates, new lending is still sluggish. There is no sufficient demand for loans and thereore nominal gdp growth is sluggish, which only creates more pessimism.


So what was the lesson learned?


  1. QE is not really money printing as such, because what we mean usually by money is the peoples bank accounts balances and cash. The portfolio effect however flattens the yield curve. So QE does not print ordinary money. However, the interest from government bond is returned to the government, so this basically is debt relief for governments.


  1. In a closed economy, the banks do not lend out their excess reserves. New lending is autonomous, i.e. new loans create new deposits and thus new money. Therefore the key for nominal growth is to create new bank deposits.


  1. In a secular staganation, and a balance sheet recession the most effective theoretical way to increase purchasing power directly is to finance government deficits outright by crediting governments’ central bank checking acconts and recognising a zero coupon perpetuity loan from the government at its face value on the balance sheet of the central bank. Most likely direct government investment is the most efficient way to increase nominal demand, although tax cuts might have a better allocationary effect.