31 Aug 2018

Market Monetarists: The Good, the Bad, and the Ugly

David Beckworth, Economics, Federal Reserve, Market Monetarism, Scott Sumner 24 Comments

Actually my title doesn’t really make sense; I couldn’t think of anything clever.

I really liked this David Beckworth post, talking about the breakdown of the Fed’s “floor system.” In fact, I liked it so much that I based my article in the forthcoming issue of The Lara-Murphy Report on it. So score +1 for the Market Monetarists.

However, in the article I had to remind the reader of conditions in the fall of 2008. I wanted to explain why the Fed was afraid to let its target rate fall further, and that’s why it “divorced” the fed funds rate from its asset purchases by instituting payment on reserves in October.

In setting the scene, I reminded the reader that crude oil prices had hit $147 a barrel in July. I started to write out that the Fed had been tightening up through the summer of 2008 because of the fear of price inflation. I thought that was the case, based on all of my reading of Scott Sumner et al. and their horror at the Fed’s focus on (price) inflation as the crisis hit.

But I wanted to be sure and get the exact figures, so I reviewed the details. And, uh, it doesn’t really look like you can explain the onset of the financial crisis by the Fed’s tight-money inflation fears:

In the chart above, the blue line shows the Fed target rate. The Fed began cutting rates back in September 2007, a full year before the financial crisis (the AIG bailout, Lehman failure, etc.). The Fed kept cutting rates, such that by May 2008, they were down to 2%–when they had been 5.25% the prior September.

Now it’s true, the Fed held rates steady from that point forward, but look at CPI. In July 2008, CPI was 5.5% higher than it had been a year earlier. (And the year/year CPI inflation rate had been near or above 4% from November 2007 onward.)

Look, I know that Scott Sumner is familiar with these figures–and I myself knew them better, back when we had just lived through this period. I’m just pointing out that at least in my case, constantly reading the Market Monetarists about how the Fed “tightened policy and causes the financial crisis” had lulled me into misremembering history.

There was a reason everybody thought Scott Sumner was crazy when they first heard his thesis, namely that Bernanke was engaged in the “tightest monetary policy since the Hoover Administration.” I think part of the reason his ideas have become more accepted, is that people have forgotten the actual sequence of events.

24 Responses to “Market Monetarists: The Good, the Bad, and the Ugly”

  1. Transformer says:

    I think Scott Sumner’s main points of reference for tight v loose money is expected NGDP and to a lesser extent expected inflation. I therefor think he would be unimpressed with your chart showing actual inflation.

    • Bob Murphy says:

      Yes I agree that Scott’s framework is internally consistent.

      • baconbacon says:

        I disagree as actually applying his framework to the 2004-2008 period would have lead one to believe that tightening was the order for the day, not loosening as the Fed had overshot all plausible inflation measures for 3-4 years prior to the crisis.

        In fact if you go from the end of the 2001 recession to the end of the 2008 recession CPI growth was almost exactly 2%, so trough to trough Sumner ought to be praising the Fed for hitting its target virtually perfectly.

        • baconbacon says:

          Further Scott’s framework isn’t internally consistent because it doesn’t define (or having read almost all of Scott’s blogposts on themoneyillusion from 2009 to around 2012 I have never seen it nailed down in an internally consistent way) how balancing is supposed to happen. If a CB over shoots a 2% target and sees 2.2% inflation for 5 years should it then attempt to under shoot by a large amount in one quarter, a lesser amount over the course of one year or attempt to mirror the overshoot and under shoot for 5 years, or undershoot by a lesser amount for 10 years.

          Further on the announcement of the CB targeting inflation or nGDP it is not clear that they should start from that date on or attempt to correct for under or overshoots in the recent or distant past. In 2012 Bernake announced that the fed would be targeting a 2% rate rather than a inflation channel. In a blog post a short time after that press conference SS had this quote discussing an old Bernake paper

          “Yes !!! Now all that remains is to ascertain the rate of NGDP growth since the second quarter of 2008: 6.1%.

          That is not a 6.1% annual growth rate, that’s the total growth in NGDP since 2008:2. And when was the last time NGDP grew that slowly over a 3 1/2 half year period? Hint: Herbert Hoover was President at the time.”

          However there is no, and never has been, an explanation for why the 2nd quarter of 2008 is the proper starting point for comparing NGDP growth or inflation rates.

          NGDP or inflation targeting demands a reference point so that it can be defined if you have been above or below your target. Without that defined point there is no internal consistency as you can claim that policy was either to tight or to loose by choosing your starting point.

  2. Andrew_FL says:

    The Fed prefers core PCE to CPI to target, don’t they?

    • Bob Murphy says:

      Yes, and even that metric was above 2% since the summer of 2007, and it was rising through the summer of 2008. So the Fed officials weren’t crazy for worrying about price inflation, even as they kept cutting rates to deal with the worsening economy.

  3. Kevin Erdmann says:

    So is the lesson you would have us take from this that (1) the Fed should have had tighter policy postures in 2008 or that (2) trailing 12 month inflation is a poor policy marker?

    • baconbacon says:

      Forward looking inflation rates don’t rescue the proposition. 5 and 10 year expected inflation rates were well above 2% for most of the period from 2004 through August 2008. According the Sumner’s position that the fed should undershoot if they have overshot and overshoot if they have undershot SS himself ought to have been calling for tightening in 2008 as the Fed had been overshooting on both forward and rear looking metrics for several years.

      Sumner’s only case that the Fed was to tight is if you start in July or August of 2008 and ignore the inflation rate/expected inflation rate prior to that.

      • Kevin Erdmann says:

        So the lesson is that both backward and forward inflation measures are poor markers for monetary policy?

        • baconbacon says:

          The lesson is that inflation targeting, using either backward or forward looking measures would not have prevented the great recession had it been done properly (as in how Sumner has described it working properly in the abstract).

    • Bob Murphy says:

      Kevin, my main lesson in this post is that Scott is wrong to say “the real problem was nominal.” I am paraphrasing of course, but I think one of his main “corrections” to the conventional narrative is to say, “It’s not that there was a ‘real crisis’ in the economy that the Fed was powerless to stop, but rather the Fed itself was the culprit with its tight money policy.”

      And I want to say no, that’s clearly not the case, as this post demonstrates. There was a “real shock” to the economy and the Fed was loosening in response to it, but its hands were tied because energy prices were skyrocketing at the same time.

      More broadly, to say the Fed had a tight policy in 2008-2009 and that’s the explanation for the Great Recession is crazy.

      • Capt. J Parker says:

        I still think Sumner has the more convincing argument. He has always acknowledged that there was a real shock to the economy from the housing market bust but this shock was not large enough to cause a Great recession. Plus, at the time, this sanguine view of the housing pullback was shared by many, including Bernanke. Here’s Sumner: “The unemployment rate barely changed from January 2006 to April 2008, despite housing construction falling in half. Even in the spring of 2008, the consensus of economists did not predict a recession, despite the fact that the subprime debacle was well understood.” From: https://www.econlib.org/archives/2015/09/how_the_subprim.html

        The Feds hands were not tied. Core inflation was not skyrocketing: https://fred.stlouisfed.org/graph/?g=l4iQ

        If I learned anything from the Great Recession it was that demand for money and increase swiftly and dramatically and that the Fed can accommodate that change in demand without causing inflation. I’ll admit that there is still a question of how to tell the difference between accommodating an increased money demand to avert a financial crisis and bubble inducing money creation. But, as for explaining the severity of the Great Recession, I think Sumner’s points are good ones.

        • Bob Murphy says:

          Hi Capt. Parker,

          I can’t devote too much time to this, but let me make two responses:

          (1) To the extent that you say Sumner’s view is compatible with Bernanke’s, then his unique perspective disappears. The “shocking” thing about Sumner is that he says “the real problem was nominal.” That’s why he says that; he is trying to jostle people who think, “Yeah, there was a real shock that the Fed should’ve done more to offset.” Sumner, to the extent that he is disagreeing with what 80% of economists were saying, has the radical position of, “No, the Fed wasn’t doing ‘too little’ to fight the Great Recession…the Fed is RESPONSIBLE for the Great Recession.” That’s certainly not what Bernanke, Krugman, DeLong… were saying.

          (2) On the housing stuff and unemployment, make sure you read this response from me.

          • Bob Murphy says:

            BTW Capt. Parker, when they revamped mises.org, the formatting on some old articles got messed up. I just re-read that article I linked for you, and the graphs are not in the right spot. You can probably figure it out, just giving a heads’ up.

          • Capt. J Parker says:

            Dr. Murphy, Thanks for the reply. I read your link and you do make a pretty case for the Great Recession being all about the housing decline in rebuttal to Sumner. But I think there is other data that tells us the timing, and in particular the scale of the housing bust does not support the idea that the housing bust by itself would cause the Great Recession. Residential construction peaked in 2006 at about $615 Billion annually. In the depths of the recession is was $225 Billion. That’s a decline of 2.6% of GDP. Okuns law says this loss of GDP from the housing bust should coincide with additional 1.3% in unemployment. Even if you allow for a big (2x) multiplier effect on GDP, the housing pullback would have resulted in at most 2.6% more unemployment. We actually got an increase of 5.5% in unemployment.
            Here is Residential Construction Spending and GDP plotted on the same scale. https://fred.stlouisfed.org/graph/?g=l5Ww I have shifted the Residential Construction Spending line up by so you can easily compare the small and slow dip in that line starting in 2006 to the big and sudden GDP dip in 2008.

            • baconbacon says:

              “But I think there is other data that tells us the timing, and in particular the scale of the housing bust does not support the idea that the housing bust by itself would cause the Great Recession.”

              The flaw of this approach is to assume that the economy was fine and that the causes of the housing boom and bust were specific to that sector of the economy. ABCT doesn’t claim that monetary expansion will cause a bubble within an otherwise healthy economy, it claims that the bubble is economy wide, but that there will be some sectors which are more inflated than others due to the way capital flows within the economy.

              • Capt. J Parker says:

                Here’s Dr. Murphy from the link he provided in response to my comment above. Dr. Murphy: “I submit that there is nothing embarrassing in the above chart for the Austrian (or Klingian) theory. As the housing boom intensified, sucking more and more workers into construction, the national unemployment rate steadily fell. Then, as the housing boom tapered off, extra workers stopped getting siphoned into the housing sector, and the national unemployment rate bottomed out. Finally, as construction employment began falling, the national unemployment rate began rising.” This doesn’t sound to me like he is talking about an economy wide bubble.

              • baconbacon says:

                @ CJP

                Its a relatively short blog post, not a thesis or a book. No blogger goes into a 4-5 paragraph summery of their core position for every post to make sure everyone is on the same page immediately.

            • baconbacon says:

              “Residential construction peaked in 2006 at about $615 Billion annually. In the depths of the recession is was $225 Billion. That’s a decline of 2.6% of GDP. Okuns law says this loss of GDP from the housing bust should coincide with additional 1.3% in unemployment”

              You can’t use Okun’s law this way, its an observed whole economy correlation not a casual relationship for fragments of the economy.

              • Capt. J Parker says:

                baconbacon, Try this: from 2005 through 2009 construction spending was very tightly correlated with construction employment. The relation is $151,000 in construction spending for each person employed in construction (my correlation is from a FRED download of TTLCONS and USCONS.) The drop in residential construction spending of $390 billion ($615B in 2005 to2.5 $225B in 2009) correlates to 2.58 million fewer construction jobs. That 1.67% of the civilian labor force. Slightly more unemployment than my Okun estimate but still no Great Recession.

              • Baconbacon says:


                I his doesn’t mean that you can extrapolate an industry to the national impact. You can estimate the total number of construction jobs lost, but that won’t tell you anything about cross sector impacts.

              • baconbacon says:

                Just to expand a little-

                If housing construction spending drops by X% and that causes a Y% decline in employment in housing construction that does not mean that the spending drop in housing only leads to that number of jobs lost in the economy. That drop will mean a decline in orders of lumber, which will mean a decline in employment in the lumber industry, it will (probably) mean a decline in homes sold which will mean a decline in the number of realtors, fewer people moving so fewer employees of moving companies, if you have large subdivisions that don’t get built then you have roads that won’t be paved, communications infrastructure that won’t get put in and retail stores that never exist.

        • baconbacon says:

          “Core inflation was not skyrocketing”

          No, core inflation wasn’t “skyrocketing” but you will note that in 2007 the inflation metric that you link was hitting its highest rates since 1994 and was above 2% for 3 straight years at that point. The Fed’s mandate in that arena is price stability and waiting for inflation to skyrocket before acting would not fulfill that obligation.

          Beyond that expected inflation (SS’s preferred metric absent a NGDP market) was rising and peaked in early July 2008 despite the economy being in recession for its 3rd straight quarter. Coincident rising UE and rising inflation and/or rising inflation expectations is a serious combination that

          “despite the fact that the subprime debacle was well understood.”

          Sumner has approvingly quoted people who dispute that the consensus on the subprime debacle is currently accurate, specifically Kevin Erdman. Further the market was demanding that AIG and Lehman post additional collateral throughout 2008 leading up to September while inflation expectations remained at 2%+. AIG’s stock price had been cut in half from May through the end of August of 2008, and 5 year inflation expectations rose and peaked in July of 2008.

          These are not signs of a market consensus, prices were much more volatile than is typically observed in a lot of areas prior to the big crash.

          SS’s thesis only holds up if you allow him to select the starting date and which metrics to care about when. The Fed had been aggressively cutting interest rates since July 2007, with the funds rate going from over 5% then to 2% in July 2008 with rising inflation expectations, and sharply dropping share prices for several key players.

  4. baconbacon says:

    I am confused about the David Beckworth post, putting my confusion here since the original post is nearly 3 weeks old and a reply to that isn’t likely to gain any traction.

    “The recent burst of Treasury bill issues has a bearing on the IOER spread over other interest rates. In particular, the greater issuance of Treasury bills should drive up (down) their yields (prices), moving other short-term interest rates in the same direction through arbitrage. This should narrow the IOER spread over other interest rates.”

    If you look at the first graph posted there are three lines which are IOR, the LIBOR overnight rate and the 1 month treasury. For most of 2011 through 2018 the IOR is the highest rate, followed by LIBOR with the 1 month treasury rate, starting around early 2018 (just eyeballing the graphs here) the 1 month Treasury basically closed the gap between it and the LIBOR rate, going from 0.1 to 0.2 percentage points lower for most of that span to pretty much the identical rate.

    My confusion is between this reality and the claim that

    “In particular, the greater issuance of Treasury bills should drive up (down) their yields (prices), moving other short-term interest rates in the same direction through arbitrage.”

    So my question is where is this arbitrage? You cannot borrow from the Treasury at the 1 month Treasury rate, you can only lend to them, so the arbitrage would have to be borrowing from one source and lending to the Treasury. But the Treasury rates are lower for almost the whole sequence, and mostly even with LIBOR for the rest, and only rarely higher. IOR is also a rate you can lend at, but not borrow from the Feds at, so you could have arbitrage between LIBOR and IOR, where you borrow from another bank at the LIBOR and then park that as reserves with the Fed, but this doesn’t have to do with the Treasury rate at all, only the spread between the Libor and IOR.

Leave a Reply