Sunday, June 21, 2015

Scott Sumner, Mark Sadowski, Jason Smith and Full Monetary Offset

     I'm glad to welcome Smith to the monetary offset wars. It's funny but this phrase by Sumner actually describes how I feel:

      "Matt also discusses a theory that I first heard from Andy Harless, that the Fed views unconventional monetary stimulus as costly, and hence fiscal stimulus might not lead to 100% offset.  This is a very logical theory, and might be true.  But it’s not as self-evident as one might assume.  For instance, the fact that the Fed has fallen short of its goals for AD might well reflect bad forecasting.  I seem to recall Yglesias pointing out that GDP growth has consistently underperformed Fed forecasts. So perhaps they tried to offset and failed.  Another possibility is that the Fed would prefer not to do monetary stimulus, and favors fiscal stimulus for that reason.  But that doesn’t mean they won’t act if necessary, just that they would prefer someone else deal with the complaints from the Ron Paul’s of the world.  There are many occasions when I hoped someone else would do something unpleasant, but when they didn’t I went ahead and did what I thought needed to be done.  Like taking out the trash."

      Indeed. Taking out the trash. This was an interesting little post where Sumner tried to explain that the Fed believes in a zero fiscal multiplier even as it says the opposite. 

      "One has to be careful interpreting the statements of Fed officials.  First of all, monetary offset can sound unpatriotic if framed one way, and not doing monetary offset can sound unpatriotic if framed another way.  Thus Fed officials sometimes say that that they take fiscal policy as a given, and do what’s best for the country given the stance of fiscal policy.  That implies offset.  They say that QE3 and forward guidance were done partly in order to offset the effects of fiscal austerity in 2013.  On the other had if you ask Ben Bernanke “If Congress does fiscal stimulus to boost employment, will you sabotage their effort with higher interest rates?” Then I’m sure he will answer no.  But that’s also the answer to that question that he would give when interest rates are positive, and even Matt Yglesias agrees that monetary offset is the right model in that case.  Framing effects.  Now of course none of this proves there is complete monetary offset, and I’ve always acknowledged that fact.  But much of the analysis of fiscal stimulus in the blogosphere, including at the highest level (i.e. Krugman) simple assumes there is no monetary offset."

    Right. When interpreting the Fed don''t listen to the Fed, listen to Scott Sumner. 

     Taking out the trash is perhaps how Sumner feels about 'Crushing the infamous thing'-of course for him the infamous thing is Keynesianism.

     When you consider the violent imagery that Sumner uses when speaking of Keynesianism-'mind-numbly stupid'-is one of the kinder things he says about it. He speaks of 'driving a stake through it's heart'; in that vein how realistic does his benefactor pleading his case sound?

    "Folks, thanks for the responses. I think I pretty much agree with all of them, and I feel like this Holy War is entirely unnecessary. I think we, and basically all economists, would agree that:

     1. in a depressed economy, fiscal stimulus can increase output if the monetary policy at the time is to ignore any effect of the fiscal stimulus (and the markets expect that monetary authority will ignore fiscal stimulus);"

     2. the monetary authority, if it wants, can always offset positive fiscal stimulus by tightening as much as the fiscal stimulus stimulates;

     3. monetary policy alone is sufficient to restore aggregate demand as long as monetary policy is not paralyzed;

     4. monetary policy can become paralyzed if the policy instrument is the nominal short-term risk-free interest rate, and that rate is at zero.

     "So why are we arguing? I just don't get it. Let's demand better monetary policy now. I think almost everyone here would agree that if the choices were (A) current policy, or (B) market-forecast-driven NGDP level targeting, B would be better. Fiscal stimulus is impossible thanks to braindead Republican deficit-obsessed obstructionism, so while I would love to see more fiscal stimulus, it's just not going to happen. Instead, let's focus on what we can do, which is better monetary policy, where no congressional cooperation is required."

     "The Federal Reserve has the wrong target. Let's fix that, and, while we're at it, put an end to the tea-leaf reading and drive policy with market forecasts. You just know it's wrong when volumes of articles are written parsing ever last word of an FOMC statement. It shouldn't matter whether Janet Yellen says "patient" or not. And with NGDPLT, it wouldn't."

      "I'm sorry that Scott and Mark are annoying everyone, but let's please keep our eye on the prize."

     Ken strikes me as a genuinely nice and reasonable guy but I don't know if he and Scott really agree on what the prize is. For him it may be getting the best demand side stabilization policy. If that's true, why does Sumner expend so much red ink on what Duda himself sees as a side issue?

    Because for Sumner, crushing Keynesianism isn't the side issue-the side issue is demand stabilization-but the main event. 

    But I know what Sumner means in doing something because no one else is doing it. For a long time I've wished some econ heavyweights would at least join the fray and call Sumner into question to the extent that he needs calling into question. 

   His theory is very intuitive but also highly problematic in a number of ways.  I'm glad that Sumner has Smith annoyed. I've been annoyed for a long time.

   As you can see, Smith literally calls Sumner and Mark Sadowski's argument here, garbage. 

   Sumner responds here.

   Anyway, we have monetary wars flaming-Simon Wren Lewis is also in it.

   and I for one couldn't be happier. It's a conversation that needs to be had. I will finish with teasing out my own views on Market Monetarism. It's composed of 2 aspects. The first aspect actually is rather disarming for Keynesian'liberal types:

   1. We should move from inflation targeting to NGDP targeting as NGDP can measure not just inflation but the kind of inflation-supply or demand side. In 2008 the Fed didn't cut rates out of worry over rising oil prices. In 2011 the ECB raised rates because of the same. NGDP would be able to differentiate. 

    I think by itself many liberals or centrists might find this attractive as many never really liked inflation targeting anyway-I am in the number. 

     Actually, at this point conservatives are a little bit skeptical-as they prefer inflation targeting- but reassurance is in the way in the shape of:

     2. Full monetary offset. We can' ever have fiscal stimulus as the Fed will offset this effect roughly 1 to 1. 

    My main issue is with 2 of course. However, even regarding 1 if we are convinced it could be a salutary change, there is the logistical question of how you know if the Fed hit its target? Wouldn''t you basically need to be able to track NGDP every single day?

    Sumner's answer is NGDP futures, an idea that till now has failed to get off the ground. No one has shown that it could work with many problems. Even Ken Duda has little to say about it. Sumner's utter silence about it recently shows it hasn't been going well. 

   Listen even if Scott were 100% right it's a debate that needs to happen. So I welcome Smith's intervention. If Sumner is right then any critique ought to actually strengthen it. The peevish and snarky way Scott usually responds to criticism suggests that it is unlikely he is anywhere near 100% right. 


  1. Hi Evilsax,

    I thought I'd respond to this comment:

    > Even Ken Duda has little to say about it.

    by at least saying something :-)

    I view that MM consists of two major ideas:

    1) Target the level-path of NGDP. That means:

    a) picking a desired level-path of future NGDP (e.g., increase 5% per year)

    b) publish the desired level-path

    c) make a firm, public commitment to hit that level-path regardless of "real" variables (unemployment, inflation, output) and also regardless of actual past NGDP (that is, do not change future targets because you missed past targets). Assert that you believe that the dual mandate is best achieved by adopting NGDP level targeting as the single, solitary goal of monetary policy. There will be no other considerations.

    d) set policy instruments (OMO's, IOR, reserve requirements, Fed Funds rate, whatever) so that your forecast of future NGDP is on target.

    2) outsource the forecasting of future NGDP as much as possible to the market. Perhaps this involves subsidizing an NGDP futures market. Or perhaps it means subsidizing a prediction market where participants make money based on the extent to which their predictions of the relationship between today's instrument settings and tomorrow's NGDP come true.

    None of this requires watching NGDP on a day-by-day basis. And, there's lots of benefit here (compared with inflation targeting) even if the NGDP forecasts the Fed targets come from Fed staff. While I think it's quite likely a prediction market would outperform Fed forecaster, I could be wrong, and even if I'm right, Fed forecasters might be good enough.

    I don't see monetary offset as relevant to the conversation. If the fiscal authority deploys massive fiscal stimulus, the Fed, under NGDPLT, would work that fact into its NGDP forecast. If the forecast indicated that the fiscal stimulus will drive NGDP over target, then the Fed would tighten (monetary offset). If the forecast indicated that we were going to undershoot our NGDP target before, but thanks to the fiscal stimulus, we are now expected to hit our target, then the Fed would not tighten and there would be no monetary offset. One could easily imagine a situation where the prediction market told the Fed that no matter what it did with its policy instruments, it was going to undershoot its target, and the Fed could then go beg congress for some fiscal stimulus, as it was truly out of ammunition. Scott asserts this case could not ever happen, that the Fed has sufficient policy instruments to hit any level of NGDP it desires. I think he's probably right, but maybe not, so isn't it wonderful that NGDPLT creates a framework within which we can assess the amount of fiscal stimulus needed to stabilize aggregate demand.

    NGDPLT is so beautiful. It appeals to my software engineer sense that the right algorithm structure can simply solve problems that are intractable when you have the wrong structure. Inflation targeting is the wrong structure. It leads to disasters like the NGDP collapse of 2008-2009. It didn't have to happen. NGDPLT would have prevented it, or at least signaled that it was coming, and the Fed would have pulled out all the stops and then gone to congress with quantitative fiscal stimulus requests.


    Kenneth Duda
    Menlo Park, CA

  2. Thank you for your very thoughtful response Ken. Listening to you I see no reason to have a Holy War over fiscal policy. Scott's not all bad-because he and I argue I never get a chance to admit he's a pretty smarty guy.

    However, one thing he lacks is an ability to build a consensus. Luckily he has you. Don't know whether you saw it but I wrote a subsequent post to this more directly about you.

    I admit I was pretty skeptical that we could get past the Holy War. Listening to you I see a plausible way forward.

    By the way, I call you Ken, feel free to call me Mike.

  3. I am now going to write a new post about your response which I will link to you after.