A James Alexander post
Jeffrey Rogers Hummel has written a very useful post about base money in the US and the difficulties in accurately tracking the number. It has led us to an attempt to explain the interaction between base money, excess reserves, interest on those reserves and the 2008 crisis.
His last three charts lead him to a conclusion with which we want to take issue. Hummel says:
Confining the definition of the monetary base and total reserves to only non-interest bearing, Fed-created outside money would yield the results for the period from 2001 to mid-2015 depicted in Figure 3, 4, and 5. With this adjustment, the mere $500 billion increase in what we can call the “outside base” since September 2008 represents merely a slightly more rapid rate of increase than the rate of increase in the base the decade prior, and nearly all of that recent increase has been in the form of hand-held currency.
No wonder that the high inflation that so many expected from quantitative easing never materialized.
Market Monetarists are clear about why high inflation has not materialized. QE has been handicapped by other elements of monetary policy. Essentially, the Fed has not wanted higher nominal growth, and it has got its wish. The Fed could be said to be anti-nominal growth and, therefore, effectively anti-growth.
In a series of posts earlier this year at Historinhas, summarised here Mark Sadowski indicates that expansion of the broad monetary base by QE Granger-caused what nominal and real economic growth the US has enjoyed. The base expansion has been nothing like enough to lead to any inflation, of course. That it is no surprise to us is due to the offsetting effect of Inflation Targeting and other restrictive monetary policies on nominal growth. The recovery the US has been, as a consequence, like pushing water uphill.
Excess reserves are commonly seen as “outside money” not “inside money”
The monetary base causes much confusion. As Hummel says, until 2008, US Base Money was essentially just “hand-held” currency, or cash in circulation, at least since WW2. There were large excess reserves built up during the Great Depression.
Hummel is wrong to classify interest-earning excess reserves (i.e. assets of the commercial banks held at the central bank) as “inside money”, at least according to Gurley and Shaw the originators of the distinction with “outside money” . The mere act of paying interest on reserves fails to change it from “outside money” to “inside money”, if one sticks to these original meanings.
“Inside money” is precisely defined. It is essentially private liabilities matched by private assets. For example, deposits at commercial banks are liabilities of the banks, but also assets of the private depositors.
Central banks control the quantity of “outside money” (ie base money)
Commercial bank deposits or reserves at the central bank are assets of the commercial banks and liabilities of the central bank, but the central bank is special. It is a public institution and as the central bank it alone can create base money (medium of account) at will, out of thin air. So, the Fed’s balance sheet grows via its own new base money creation. Very broadly speaking, commercial banks balance sheets can only grow in aggregate if the Fed allows base money to grow. It has become harder to track this phenomenon in recent years with the rise of financial disintermediation.
The whole “inside/outside money” concept is not commonly used by monetary economists who tend to avoid using the terms. They are more familiar to specialists in finance. Much of the confusion arises because money for monetary economists is “outside money” or base money, and is narrower than money in the mind of the general public.
Not all money is the same, but not quite how Hummel sees the distinction
Hummel’s “outside base” money concept could be even narrower, but is not in common usage. Changes in base money are what predict changes in the nominal economy, which Market Monetarists see as sometimes having heavy impacts on the real economy in the short-term and medium-term. Changes in these broader definitions of money have much less predictive power. It is best not to think of “inside money”, or non-base money, as money at all when it comes to thinking about monetary policy.
The velocity of the commercial bank reserves with the central bank can be just as volatile as velocity of currency in circulation. Both types of money can be hoarded during negative shocks, and rendered inert, other things being equal. A negative shock drives up the value of money and drives down the prices of non-money (deflation). Or vice versa, a positive shock, i.e. “hot economy”, drives down the value of money and drives up the prices of non-money (inflation).
Interest on excess reserves would be unimportant above the ZLB
More importantly, whether these excess reserves earn interest or not is almost an accident of history. The Fed could have chosen to pay zero or negative rates on these reserves if it wanted. In the circumstances, the decision to pay interest on excess reserves had some profound consequences but not the ones Hummel brings up. As Peter Ireland wrote in his widely referenced article on excess reserves:
the Federal Reserve’s [then] recent decision to begin paying interest on reserves is unlikely to have large effects on the behavior of macroeconomic variables such as aggregate output and inflation, once normal times return. [my emphasis]
There were hardly any excess reserves in the post-WW2 era. Why would there be? The interest rate on excess reserves was zero and market rates were in the hundreds of basis points. There were also no serious, system-wide, bank runs.
The financial crisis changed everything. Commercial bank (excess) reserves at the central bank exploded when Lehman went bust. These reserves were created as the inevitable counterpart of the emergency Treasury/Federal Reserve funding of banks and other financial institutions who’d found that in the pre-Lehman, but especially post-Lehman, liquidity crisis market-sourced funds were hard to come by.
The trouble with IOER, explained better
The decision to pay interest on those reserves highlights a somewhat hidden episode of the financial crisis, but one that Scott Sumner , David Beckworth and others have repeatedly called into question.
IOER was a new concept for the US but is actually a very familiar part of the monetary policy landscape in the majority of monetary regimes. It is a second tool with which central banks can control the monetary base, by altering demand from banks for money, in addition to traditional open market operations buying financials assets (with new money, like QE) or selling financial assets (for existing money and thus reducing the amount of base money). In most regimes IOER is a little below market rates and has no significant impact on monetary policy or the amount of base money as it is more attractive to buy other short-term risk-free assets like T-bills.
Most countries also have required reserves, usually as a percentage of some definition of deposits. In the US they are currently $91bn, versus $2,500bn of excess reserves or 25x as much.
However, it is not the existence of IOER that is the problem, but the level of the remuneration in an environment of zero or near zero interest rates. The Federal Reserve decided to pay a rate of IOER above overnight and T-Bill rates because the market was driving those rates below where the Fed thought was desirable.
The crisis-induced birth of IOER – not a good omen
Our chart shows how the market, as represented by the 4-Week T-Bill was constantly below overnight rates in the run up to the crisis, well below both the US (Effective Fed Funds rate) and London (LIBOR) money market rates. The “money market” is something of a misnomer as the “market” element is heavily influenced, if not actually controlled, by what target interest rate the central bank desires.
The liquidity squeezes caused by financial turmoil in August/September 2007, March 2008 and then the third and final time in September 2008 are clearly visible by high levels of volatility. The fact that the 4-Week rate often fell further than rate cuts on rate cuts days shows how the market constantly worried that the Fed was behind the curve. The collapse in the 4-Week rate on the morning after the Lehman failure should have been the signal for the Fed to cut rates, as that was what the market both expected and wanted. But the Fed delayed, causing turmoil in money markets, that even spread to the rate closest to its control, the Effective Fed Funds.
The turmoil in money markets seen in the chart happened because the Fed was not doing enough to cope with surging demand for money, by supplying more. And what money there was available had to be hoarded in some other risk-free place, causing the collapsing T-Bill rates as their prices were bid up, especially in times of stress.
Some non-bank market participants had direct access to them, but most non-bank economic actors just moved their money to the biggest, safest banks. Because those banks had no need for the deposits, in fact they wanted to shrink the risks on their balance sheets, the biggest, safest banks also bought T-Bills. Smaller banks were also de-risking their balance sheets, increasing demand for T-Bills too. And traders were obviously buying T-Bills in anticipation of the Fed cutting its target rate.
Was the Federal Reserve right to ignore what the markets were expecting and wanting, rates at zero and/or lower, or a massive injection of liquidity? To be fair it wasn’t just Fed fear of negative rates. Other pressures were bearing down on the Fed. For instance, going below zero would cause many money market funds to “break the buck”, i.e. break an implicit promise to protect investors capital and thus mask how they were sold, as higher remunerated, but unsecured, deposits. This prompted further concern about instability, illustrating that moral hazard was not just confined to banks, of course.
But by instituting IOER the Fed actually added to the demand for money, as banks too in their own operations moved from higher risk non-central bank assets, to zero risk central bank reserves. This was not helpful in a time of liquidity crisis.
Even with IOER from early October 2008, 4-week T-bill rates still went negative twice during December 2008. T-bills have also still been negative on at least 36 days since 2009 and have only averaged a measly six basis points over that period, showing just how strong demand for money remains versus supply and thus how tight monetary policy remains.
It’s not just about M but V too
“What if” there had been no IOER is harder to establish. Market pressures would probably have driven those rates negative, just like they have in many parts of Europe in recent years. This should lead to a quicker recovery other things being equal. But it is hard to forecast how negative rates in September 2008 would have played out, if money market funds had seen a flight too.
New money would still need to have been created to cope with the increased demand for it. As Mark Sadowski has shown, approximately each 10% rise in base money leads to a 1% rise in NGDP.
But that is only the money quantity part of MV working. The key to a recovery caused by a large drop in V is to raise expectations of more M, not just more M itself. If it is just more M it is not certain that the recovery will become self-sustaining unless there is a belief that the extra M is either permanent or will only be withdrawn once V is back up and running, i.e. after confidence has been restored.
The Federal Reserve has signally failed to engender that belief in increased M being permanent: it constantly, and very publicly, plans for a reduction in the balance sheet – causing uncertainty. It doesn’t engender confidence in growth as it constantly and very publicly searches for traces of inflation picking up where none exist. And it constantly, and very publicly, supports the Philips Curve/Slack Theory of macro and so for the last several years has led the market to expect a break out of wage inflation and immediate monetary tightening. Thanks for that!
But worse, it is so self-defeating. The constant and public hand wringing has led directly to the very sluggish recovery, and the constant and very public threat of a near term rate rise now appears to be slowing even this. Thanks again!
No monetisation, just poor, confusing monetary policy
Hummel also says this:
Thus, at least $2.5 trillion of the post-crisis explosion of the monetary base constitutes interest-bearing inside money that in substance is government debt merely intermediated by the Fed.
This is wrong, government debt has not been monetised. The Treasury intends to repay it. Government debt was bought because it was available. The Fed also bought hundreds of billions of agency debt, funding US housing. If the market really believed all this government debt and mortgage bonds had been monetised then inflation really would have been a problem. But as Hummel has noticed, it hasn’t been a problem at all. In fact, the opposite. The Fed has made it clear it will sell back the debt over time. And the market believes it. Some monetisation!
Perhaps some monetisation might have been a good thing and engendered a quicker recovery.