Wednesday, October 4, 2017

The confusion between Quantitative Easing and Interest on Reserves

Arnold Kling has a link to some commentary on QE from economists at First Trust.  They question whether QE was effective.  I agree with some of their comments.  I do think the issues with mark-to-mark accounting and some of the other actions taken in spring 2009 did contribute to the turnaround.  But, I think we need to be careful about exactly how we identify Fed policies during that period.

The problem is that in October 2008, the Fed initiated interest on reserves in a strange and misguided attempt to keep the fed funds rate at 2% even though the Fed clearly should have been pumping cash into the economy, which would have pushed the fed funds rate lower.  For most of the next two months, the interest on reserves rate was actually higher than the effective federal funds rate even though it was lower than their stated target rate.  In other words, Fed policy was so tight at the time that they would have basically had to sell their entire remaining stock of treasuries to meet their target rate, sucking cash out of the economy.  They couldn't do that, so they used interest on reserves as a sort of back door way to obtain cash.  Instead of pulling currency out of the economy, they borrowed cash from the banks.  They induced banks to lend to them instead of lending into the economy.  They bought treasuries with that cash, so another way to describe what happened is that they induced banks to lend to the government instead of to private parties, and they acted as an intermediary, with excess reserves as the bridge between banks and treasuries.

This was a contractionary policy, and because they were acting as an intermediary between the banks and the treasury, the Fed's balance sheet ballooned.

Then, at the end of 2008, and later in 2010 and 2012, they did three rounds of quantitative easing.  Since by then rates were near zero, this meant that they injected cash into the economy without using the framing device of raising or lowering rates.  The use of interest rates as the framing device for describing monetary policy is strange anyway.  I wish they didn't do that.  Since the neutral rate is a moving target, it leads to a lot of misguided opinions about whether policy is loose or tight.  And, it causes the idea that monetary policy works by making cheap debt financing available to be greatly inflated in the minds of the public.  It seems that most of the commentary, even among financial experts, about monetary policy and the business cycle is framed that way, and I think that is very damaging.  It would be better to base financial analysis on astrology.  At least that would lead to an unbiased random outcome of analysis.  It's like the experiment where you put a moth and a fly in two separate open jars with the bottoms facing a light.  The putatively smarter moth keeps knocking himself against the bottom of the jar in an attempt to escape toward the light while the fly just flies around randomly until he escapes the jar.

Anyway, it is odd that QE is framed as some sort of new type of policy.  It was August 2007 to September 2008 where the Fed had a strange new policy.  They were making a series of emergency loans to failing financial firms, but they were "sterilizing" those loans by selling treasuries into the market, so that those emergency loans didn't cause currency to expand.  When the Fed initiated QE at the end of 2008, that was actually a return to normal monetary policy.  That was the first time in more than a year that they actually just bought some treasuries in open market operations with the motive of injecting cash into the economy in order to create inflation.

This was an accommodative policy, and because interest rates were near zero, the incentive of private institutions to avoid holding cash was low, that cash was simply held at banks, and the Fed's balance sheet ballooned.

We can see these effects in this Fred graph of the Fed balance sheet and 5 year inflation expectations.  In late 2008, inflation expectations cratered when interest on reserves was initiated.  Policy was highly contractionary and the balance sheet bloated.

For the first part of QE1, the balance sheet didn't actually expand, because while QE1 was expanding the balance sheet, the initial surge of excess reserves when the Fed pegged the short term rate target at 2% was now unwinding, since the rate was near zero.  By mid 2009, QE1 was growing the balance sheet.

Then QE1 was ended, but QE2 began in late 2010, and we see the balance sheet grow again.

Now, to me, just eyeballing inflation expectations, I see expected inflation rise as the initial interest on reserves policy gives way to QE1.  Then, it recedes after QE1 ends, then it rises again when QE2 begins, and recedes again as QE2 ends.  There is little reaction to QE3, but when QE3 ends, inflation expectations again recede.  The QEs seem to be weakly effective.

I have also included the nominal 10 year treasury rate in the graph.  It gives a more clear signal with the QEs, rising each time and declining after they end - even during QE3.  Each time, there is a little hump in 10 year rates.  This is an example of how thinking of monetary policy in terms of the cost of credit is strange.  I would consider the fact that 10 year treasury rates increased during the QEs to be a sign of the success of the QEs.  That success has nothing to do with cheap credit.  In fact, credit became more dear because expectations and investment demand improved.  If only the Fed had continued QE persistently instead of running it in starts and stops.

But, seeing the growing balance sheet as some sort of unified invitation to hyperinflation is incorrect.  First, the QEs were effective, but only weakly so, so it would take a lot of asset growth to trigger a little inflation.  Secondly, the initial growth in the Fed balance sheet was actually contractionary.


  1. Egads!

    There are as many explanations of what really is/was QE, and what effects it really has/had, as there are macroeconomists.

    Maybe more explanations than macroeconomists--some observers have had evolving definitions of QE, when foretold results never materialized (runaway inflation, for example).

    One of the great ongoing acts in intellectual vaudeville is macroeconomists asserting, with great certitude, their descriptions of QE, which are diametrically opposed to other macroeconomist's assertions, also delivered with equally held assurance.

    And it ain't over yet. Now we have declarations that if the Fed does not shrink its balance sheet, we will have (you guessed it) runaway inflation.

    Oddly, there is no inflation in Japan, and they are continuously adding to the Bank of Japan's balance sheet and have no plans to sell the balance sheet ever, from what I can tell.

    Then we have the view QE would have been effective, except for IOER.

    Or David Beckworth's view, that the Fed needed to say QE was permanent to be effective.

    I always ask, "If QE is inert, why not have the Fed buy a few more trillion in Treasuries, and thus relieve taxpayers?"
    I get wrinkled noses, but no real answers.

    Then there is the mysterious Primary Dealers Credit Facility.

    You know the Fed does not buy bonds on the open market?

    The Fed buys bonds from 21 primary dealers. The dealers buy the bonds from customers and sell the bonds to the Fed, and the Fed credits the primary dealers commercial bank accounts by an equal amount.

    Except the Fed loaned money to the primary dealers to buy the bonds, back in 2008.

    With great certitude, I asset my view: The Fed printed (digitized) money in QE, and bought bonds with it. The primary dealers are only an intermediary, forget them.

    The sellers of the bonds into QE then had $4 trillion in digital cash. They bought other bonds, stocks, property, deposited the money or converted digital to paper cash.

    This is somewhat stimulative. Maybe should have been bigger.

    But remember, there are global bond markets, and worldwide there are more than $100 trillion in bonds outstanding and more than $210 trillion in debt. The Fed actions, given global markets, seem small.

    Helicopter drops are probably a better idea.

  2. The sellers of the bonds into QE then had $4 trillion in digital cash. They bought other bonds, stocks, property, deposited the money or converted digital to paper cash.


    I should have added, "or spent the money."

    This seems important to me. When the Fed buys bonds, no one in the private-sector has to give up present investment to consumption.

    When I sell a T-bond to Kevin Erdmann, he has to give up other investments, or consumption.

    When I sell a bond to the Fed, I can consume immediately, but no one else has to refrain.