Tuesday, 24 February 2015

Monetary offset is a joke.

One of the central ideas, if not THE CENTRAL idea of Market Monetarism is so called “monetary offset”. Scott Sumner is one of the World’s leading proponents of monetary offset, if not THE leading proponent. He explains the idea in an article entitled “Why the Fiscal Multiplier is Roughly Zero”.

Monetary offset according to Sumner is the idea that fiscal stimulus is very ineffective (as the title of his article implies) and THE REASON apparently is that if the fiscal authorities implement too much stimulus, the monetary authorities (i.e. the central bank) will “adopt a more contractionary monetary policy in order to prevent inflation from exceeding their 2 percent target.”

Now anyone with a grasp of economics ought to be able to spot the flaw in that idea. Incidentally I’ve put the relevant passage from Sumner’s article below under the heading “Sumner’s own words” and in italics. And of course readers wanting an even more detailed look at his ideas are free to read his whole article.

Anyway, for the benefit of readers who haven’t spotted the flaw, I’ll spell it out, and in fact the flaw can be illustrated very nicely by reference to a car, as follows.

Suppose one person has control of the accelerator (fiscal policy), and someone else controls the brakes (monetary policy), obviously one of the things the “brake controller” will do is to apply the brakes if the “accelerator controller” has stepped on the gas too much and the car is exceeding the speed limit.

Sumner’s conclusion from the latter is that the accelerator (fiscal policy) is near useless, because if too much of it is applied, the brake controller will slow down the car. Well hopefully most readers will by now have seen the flaw in the argument.

The flaw of course is that an accelerators is a good way of controlling a car’s speed. There’s nothing inherently wrong with accelerators. And the fact that drivers sometimes to too fast and need to apply the brakes is not an argument against accelerators.

Moreover, the idea that if inflation looks like getting excessive, that the central bank will apply the brakes is not exactly an original idea. Everyone including the average taxi driver knows central banks do that. I.e. there is no need whatever or a gradiouse new theory called “monetary offset”.

Monetary offset is nothing more than a verbal sleight of hand. It’s for people who don’t like fiscal policy, but can’t find any serious flaws in fiscal policy.

Sumner’s own words.

“Why has the effect of fiscal stimulus been so meager in recent years? After all, interest rates in the United States have been close to zero since the end of 2008. The most likely explanation is monetary offset, a concept built into modern central bank policy but poorly understood. We can visualize monetary offset with the Keynesian aggregate supply and demand diagram used in introductory economics textbooks. If fiscal stimulus works, it’s by shifting the aggregate demand (AD) curve to the right. This tends to raise both prices and output as the economy moves from point A to point B, although in the very long run, only prices are affected. Now let’s assume that the central bank is targeting inflation at 2 percent. If fiscal stimulus shifts the AD curve to the right, then prices will tend to rise. The central bank then must adopt a more contractionary monetary policy in order to prevent inflation from exceeding their 2 percent target. The contractionary monetary policy shifts AD back to the left, offsetting the effect of the fiscal stimulus. This is called monetary offset.”


  1. I accidentally deleted a comment by Ken Duda. Here is his comment:

    Ralph, this is a bad post for two reasons.

    Let's divide economists into three groups: Austrians (who prefer morality to reality), Neoclassicals (who prefer math to reality), and Reasonables (everyone else, from Krugman to Cochrane with lots more in between).

    I think all reasonables agree that monetary offset is possible. It is not "a joke". If the Fed wants, it can draw all of the money out of the economy, which will overwhelm any reasonable amount of fiscal stimulus. Or it can cause hyperinflation. Likewise, if the fiscal authority wants, it can trash the economy in either direction, either by raising tax rates to 200% (kill everything off) or by cutting them to -200% (hyperinflation). Both monetary and fiscal policy are potent. So your argument that monetary offset is "a joke" is wrong on the face of it.

    The deeper sense in which you are wrong is I believe that basically any reasonable economist (as defined above) would agree that there is a new consensus policy is a huge improvement over current policy, specifically:

    1. replace IT with NGDPLT at the Fed
    2. introduce automatic fiscal stabilizers (tax credits) if NGDP drops too far below target for too long

    I think every reasonable economist of whatever flavor (Keynesian, MM'er, anything in between) should agree that the above is a dramatic improvement over current policy.

    By taking an unreasonably extreme position against fiscal stimulus, MM'ers are making their idea of "perfect" the enemy of the good. And you are doing the same thing here. By calling monetary offset "a joke", you are implying that monetary policy does not matter. You aren't saying that, but you're implying it, and it's the wrong thing to imply, because it's not true, and it argues against NGDPLT, which would be a massive improvement in monetary policy even if it doesn't solve all problems. All responsible economists should remind their readers every chance they get that the Monetary/Fiscal Consensus would be a dramatic improvement over current policy and we should all push for its immediate adoption, for the sake of millions of future employed people.


    Kenneth Duda
    Menlo Park, CA

    1. Kenneth,

      First, I like your definitions of Austrians and Neoclassicals. Moving on to your next para (starting “I think..”), I quite agree that offset “is possible” in the sense that the central bank can negate any fiscal action. I said as much in the above article. I also agree that both fiscal and monetary policies are “potent” as you put it.

      But that doesn’t destroy my basic point, which is that it's nonsense to argue that because it’s possible for A to mess up B that therefor B is useless, which is what Scott Sumner is saying. I can mess up a computer by hitting it with a hammer. That doesn’t prove computers are useless.

      Next, and re NGDPL, that is a SEPARATE issue to the “fiscal policy is useless or useful” debate. As I pointed out once on David Beckworth’s blog, assuming both fiscal and monetary stimulus work, one can target NGDP (or target inflation) with either fiscal or monetary stimulus. As I remember, he didn’t disagree with me.

      My views on NGDP are not very relevant here, nor am I an expert on it, but I don’t have any big objections to it. I’d just make the point that the Bank of England has been described as a “closet NGDP targetter”, which seems to be accurate. E.g. at the start of the recent recession, the BoE cut interest rates DESPITE inflation being above target because they thought inflation was largely cost push.

      Next, re your advocacy of “automatic fiscal stabilisers”, that’s not consistent with Sumner’s idea that fiscal stimulus is near useless. But if you’re saying we should have a mix of monetary and fiscal policy, then I quite agree. I actually favor aiming for exact dollar for dollar mix: i.e. feed one extra dollar of base money into the economy for every dollar increase in public spending (or tax cuts).

      That amounts to simply having the state create new money and spend it (or cut taxes) in a recession. And if anyone wants to use that “mix” to target NGDP rather than inflation, I have no objections. (Incidentally I would’nt insist on an absolute cast iron dollar for dollar equality, there. But I’d favor something approaching equality).

      Finally, you say: “By calling monetary offset "a joke", you are implying that monetary policy does not matter.” Not at all. I fully accept that “money matters” (to use an out of date phrase). That’s why I favour, as just stated above, feeding new money into the economy in a recession.

      To repeat, the “joke” I’m attacking is the argument that because A can be used to negate or mess up B, that therefor B is useless.

    2. "But that doesn’t destroy my basic point, which is that it's nonsense to argue that because it’s possible for A to mess up B that therefor B is useless, which is what Scott Sumner is saying."

      That is not what Sumner is arguing. He is not only arguing that it is "possible for A to mess up B" but that the central bank's too low inflation target is a standing promise to mess up B.

      He is arguing that it is futile to try fiscal stimulation when the central bank promises to undo all its benefits and has the power to do so.

      He is also not arguing that B can't work . He is arguing that once you remove the central bank's promise to strangulate the economy, the fiscal stimulation might work but it probably is no longer necessary because removing the promise not only allows fiscal stimulation to work but allows private markets to resolve demand problems by themselves.

      Government fiscal spending is definitely not Sumner's preferred method to solve demand problems but he has not stated that it can't work once you remove monetary barriers, only that it would probably no longer be necessary.

    3. Benoit,

      Thanks for that, and I think you’ve certainly spotted a weakness in my argument. However I’m not sure about your claim that Sumner thinks the Fed’s inflation target is “too low”. Sumner doesn’t say anything about it being too low in the Mercatus article dealt with above. However I’m aware that he favors NGDP targetting, and that might involve going above the 2% target from time to time.

      But this is too complicated to deal with in a comment after a blog article. I’m going to have to go thru that Mercatus article in detail and re-write my criticisms of it in a new blog post. I’ll put that on my list of upcoming blog posts, and put it online in due course.

    4. Thanks for the reply Ralph.

      Here is a post where Sumner advocates for a 4% inflation target as second best solution (even if it is just as a compromise with Keynesians):


    5. Benoit,

      On re-reading this 8 months later….

      You say “..removing the promise (by the central bank not to mess up the economy) not only allows fiscal stimulation to work but allows private markets to resolve demand problems by themselves.”

      Where on Earth does Sumner make that “demand” point? If Sumner has some new theory about the market automatically “resolving demand problems”, I’m all ears. One of Keynes’s basic points (and I think he was right) was that free markets DO NOT automatically deal with deficient demand – at least they certainly do not deal with deficient demand nearly as quickly as would be desirable.

      Next, you say “He is not only arguing that it is "possible for A to mess up B…”. So you agree with me that one of the things he DOES say is that it is possible for A to mess up B, ergo B is no use (which is, as I point out above, an illogical argument)?

      Next, re Sumner’s advocacy of a 4% inflation target, that’s a very unimportant point. First, I don’t think it makes a HUGE difference whether inflation is 1%, 2% or 4%. Second, as you say, Sumner says that a 4% target is a “second best” solution. I agree.

  2. Thank you Ralph.

    Not surprisingly, we fully agree on the substance here.


  3. Even if you relax the assumptions in mainstream models, the idea of an offset sort of makes sense. If the central bank is following a feedback control law to stabilise the economy, and you assume that monetary policy is effective, fiscal policy would be cancelled out by monetary policy.

    But there are two problems.

    To parallel what you wrote, if stimulus is needed and the fiscal authority acts first, the central bank does not need to apply stimulus. You end up in the same place, but it was fiscal policy that acted.

    The bigger problem with Sumner's argument is the theory that monetary policy was tight since the recession. You would be hard pressed to find anyone who would argue that, other than those who argue that monetary policy is ineffective at the zero bound. Since Sumner has some theory why monetary policy is effective at the zero bound, none of this makes any sense.

  4. Brian... what?

    Monetary policy has been too tight, from 2008 to present. You can see this in two ways: NGDP growth (still below trend), and long-term interest rates (still at historic lows). Long-term rates are low because the market is betting on an extended depression because money is too tight. Whether you believe that the Fed can't loosen money further as a technical matter (Keynesian), or whether you believe the problem is a lack of proper policy (Market Monetarists), I think nearly all mainstream economists agree money is too tight. They just say it differently. Keynesians say, "conventional policy is up against the ZLB, the Fed should try to do more, we're not sure it can." Market Monetarists say, "money is too tight."

    The only people saying money is not tight are people who think that low interest rates is the same thing as loose money. During the 1970's, money was too loose, right? And interest rates were what, 6 to 10%?

    Sumner (and I) believe that the right monetary policy (such as NGDP level targeting) would be effective at the zero lower bound, and could offset fiscal policy even at the zero lower bound. This belief has theoretical basis but is untested as a practical matter. I realize there is risk that it is wrong. Thus, I believe the best policy to stabilize the economy consists of two things:

    1. NGDPLT
    2. automatic fiscal stabilizers that kick in if NGDP falls below target by too much for too long



Post a comment.