What is demand stimulus?

This is a sort of follow-up to my previous post.  One can think of demand stimulus as policies that boost NGDP.  (There are of course other policies that boost RGDP, such as supply side reforms, which work even if NGDP doesn’t rise.  But demand stimulus boosts NGDP.)

We know from long run money neutrality that the long run trend rate of growth doesn’t matter, except for second order effects like hysteresis and menu costs and taxation of capital income—and these second order effects might be positive or negative.  If someone argues that a certain policy may be able to significantly raise the trend line for RGDP, they may be right, but they are almost certainly NOT talking about demand-side stimulus.

The upshot of all of this is that there is only one coherent way to think about demand-side policies.  When should AD be more expansionary than average and when should it be less expansionary than average? It’s incoherent to say, “I think demand side polices should always be stimulative.”  That doesn’t even mean anything.  It’s like saying, “I believe all Americans should earn above average incomes.”  Any demand-side strategy should either call for stable AD growth, or else specify when aggregate demand should be more expansionary than average and when it should be more contractionary than average.

If you are advocating demand stimulus during a period of low unemployment, then (whether you know this or not) you are implicitly suggesting that demand-side policy should be more contractionary than average during a recession.  Not good.

A corollary of this is that terms like ‘hawks’ and ‘doves’ don’t have the meaning that almost everyone thinks they have.  If you have a 2% inflation target, exactly how do you implement a “dovish” policy?  A “hawkish” policy?

What if we turn to fiscal policy; does that change things?  Not at all.  The government’s national debt is constrained by the fact that the debt must be serviced in the long run.  This budget constraint means that budget deficits that are larger than average during certain periods must be offset by deficits that are smaller than average during other periods–to keep the debt manageable.  It makes no sense for someone to say, “I generally favor a more expansionary fiscal policy than what is favored by Sumner.”  It’s not even a coherent statement.  If you say that you favor a more expansionary fiscal policy that what I currently favor, you are implicitly saying, “and at some future date I prefer a more contractionary fiscal policy than what Sumner will favor at that point in time.”  I worry that the insights of Robert Lucas are being forgotten.



30 Responses to “What is demand stimulus?”

  1. Gravatar of Tyler Cowen Tyler Cowen
    16. February 2018 at 12:47

    If you think there is lots of output restriction today, it is easy enough to pull it out with some fiscal policy right now, and then later pay off those debts *with money creation*. The later policy doesn’t have to be contractionary, and it might be fully neutral with respect to monopoly, but still output has gone up overall.

  2. Gravatar of Student Student
    16. February 2018 at 13:07

    Tyler’s comment is what I was trying to say with my comment on the previous post, I just did so poorly.

  3. Gravatar of Philo Philo
    16. February 2018 at 14:00

    Is this a lexicographical remark, about the meaning of ‘expansionary’ and ‘contractionary’–namely, that ‘expansionary’ means ‘more stimulative than average’ while ‘contractionary’ means ‘less stimulative than average’? Such a lexicographical claim would be dubious: “average” *over what time period*? Besides, suppose that running 1% deficits on average made the real economy grow faster over the long run than running a balanced budget on average: then it would seem natural to call the former policy–a long-term policy rather than something temporary–“expansionary” *in comparison with the latter*. (Maybe this supposition–faster real growth with deficits on average–is contrary to somebody’s economic theory, but is it proper to build that theory into the definitions of ‘expansionary’ and ‘contractionary’?)

  4. Gravatar of ssumner ssumner
    16. February 2018 at 14:53

    Tyler, I don’t think you can approach these issues in that way, you need to think in terms of a long run regime. Even if you try to combine monetary and fiscal policy in the way you suggest (by monetizing debts), you end up with policy regime where policy is more expansionary than average some periods, and less expansionary that average in other periods. And Friedman’s (1968) great insight was that a policy that is less expansionary than average is actually contractionary. AFAIK, that’s the implication of all serious modern macro models. There’s no way around that implication, without rejecting long run neutrality of nominal changes. (Again supply side effects are different, maybe infrastructure boosts productivity, for instance.)

    People made fun of Krugman saying that reducing the deficit from 1 trillion to 500 billion was contractionary, but that’s actually the implication of the standard AS/AD approach to macro (assuming 500 billion is below average, with is perhaps questionable). If policy could be consistently expansionary then it would seem to violate the long run neutrality of money, allowing Bangladesh to get rich by printing money or running deficits (not saying you believe that, just pointing out what seems to me to be an implication of thinking in terms of expansionary demand side policies that are not offset by subsequent contractionary policies.)

    Philo, You said:

    “Is this a lexicographical remark, about the meaning of ‘expansionary’ and ‘contractionary’–namely, that ‘expansionary’ means ‘more stimulative than average’ while ‘contractionary’ means ‘less stimulative than average’?”

    We talk that way because macro models suggest that demand stimulus is only expansionary for RGDP to the extent it is more expansionary than average, and is contractionary to the extent it is less expansionary than average. You might not like the terminology, but there are sound theoretical reasons for making that distinction.

    Paul Volcker’s 4% inflation policy of 1982-83 was contractionary at the time. That policy today would be expansionary, as it would represent an increase in inflation.

  5. Gravatar of doug M doug M
    16. February 2018 at 15:03

    If you have a 2% inflation target, exactly how do you implement a “dovish” policy? A “hawkish” policy?

    You are a hawk or a dove in terms of how you interpret that policy.

    Is 2% as ceiling an average or a floor?

    If you are above target, or below target, how aggressively should any course corrections be? Should the course corrections be more aggressive depending on whether the economy is on the “boom side” or the “bust side” of the growth trend?

  6. Gravatar of BC BC
    16. February 2018 at 15:44

    I think of the dove and hawk designations as denoting the bias or errors that one makes in implementing discretionary policy. Doves tend to underestimate the likelihood that low inflation is “transitory” and overestimate the likelihood that high inflation is transitory and thus tend to overestimate the amount of stimulus needed when inflation is low and underestimate the amount of contractionary policy needed when inflation is high. Vice versa for hawks. One can also think in terms of expected future inflation, which is unobservable. Doves’ inflation expectations are persistently lower than hawks’ expectations. Thus, under the same current conditions, doves tend to advocate more stimulus than hawks.

    Imagine driving a car when fog limits one’s visability. Both doves and hawks agree that they should stay in the middle of the road on average. However, doves believe that the road is going to turn to the left while hawks believe it will turn to the right.

  7. Gravatar of B Cole B Cole
    16. February 2018 at 16:51

    Well I agree with this post…but.

    If the US runs into another patch of inflation—say prices double in 10 years—then the national debt shrinks to GDP, even if modest deficits persist.

    And then Japan. The Bank of Japan now owns 45% of JGBs outstanding and they cannot hit even 1% inflation. The national budget is chronic deficit. Are monetary and fiscal policy the same thing in Japan?

    And are contractionary policies truly neutral in the long run?

    If an economy can only expand in relation to its base then a contractionary policy can have permanent ramifications.

    On a practical level, we saw a 20-year period of stagnation in Japan caused by contractionary policies. We saw the Great Depression in the US.

    How many lifetimes do you have to live awaiting neutrality?

  8. Gravatar of dtoh dtoh
    16. February 2018 at 17:08


    So here are some questions. 🙂

    Assume first that the Fed policy is based on an NGDP target and they stick to it (i.e. don’t change the target) and are competent.

    1. Over the long run, will RGDP growth be the same with a 0% NGDP target as with a 10% NGDP target?

    2. Same question for the short run and the medium run.

    3. If there is a short term change in RGDP growth which is dependent on the NGDP target, will we still nevertheless end up with identical output in the long term. I.e. regardless of whether the Fed sets a 0% target or 10% target today and sticks to it, will output be identical in 25 years?

    4. If you answer yes to Question 1, does that mean sticky wages and prices are learned behavior and will go away with the right policies?

    5. If you answer yes to Question 4, does that mean you’re an Austrian and agree with Major Freedom?

  9. Gravatar of Bob Bob
    16. February 2018 at 20:08

    The people that disagree with your outcomes will just disagree with your basic premises: Maybe Tyler’s way, by adding more variables: Just like you’d disagree with someone that thought that monetary policy is not really a thing.

    MMTers would also disagree, and say that debt never needs to be serviced, and that if you do you are just punishing the private sector. I don’t think they are right, but it’s a way out.

    While I suspect that there are more useful variables than the ones you discuss, I don’t think they really apply all that well today, so you are right, especially at this point in time.

    Ultimately though, I don’t think that this argument of yours will be all that effective: It’s not as if the argument for expansionary fiscal policy today goes any deeper than ‘Expansionary policies help the party in power, and therefore if one likes the current regime, then one should favor expansionary policy regardless’. Ultimately, deeper arguments are excuses, not unlike states rights vs federal power.

  10. Gravatar of Benjamin Cole Benjamin Cole
    16. February 2018 at 21:36

    OT but in the ballpark

    I get a sense that many regard national borrowing as a bit of a morality play. That is a compelling sentiment for me too, btw. Moral hazard. One should pay back what has been borrowed, three little piggies and all of that. I am in.

    But is paying back national debt “right” in terms of macroeconomic policy?

    The Greeks must pay back what they borrowed. Moral hazard. Egads, we have crushed the Greek economy for 10 years straight, no end in sight. It appears Greek citizens will live out their lives at 25% lower living standards than before. Don’t believe me, see this:


    Is what has happened to Greece a reasonable macroeconomic policy?

    Back in the US, from here, I contend the question is not moral hazard, but what we do. We have sunk that low. Congress is going to borrow $1 trillion a year for the foreseeable future. Right or wrong has nothing to do with it.

    How do we pay down the debt without doing a Greece?

    Don’t answer, “Oh, don’t borrow as much.” That solution is not on the table.

    Japan is doing regular QE, and it seems to work.

    We see the Fed often goes to prophylactic interest rate hikes. Right now, for example. Inflation is well under target, but the Fed is raising rates, citing the potential for inflation to rise.

    Should the Fed engage in prophylactic QE? If ever nominal GDP growth gets below a level, that triggers QE. A Taylor-type rule requiring QE? This would have the added benefit of cutting the national debt. The Fed would funnel interest rate payments back to the Treasury. The federal government would essentially own the debt it issued, a type of mobius-strip economics.

    Maybe we do a Greece. Maybe we do an Argentina.

    Is a better option, maybe we do a Japan?

  11. Gravatar of Jerry Brown Jerry Brown
    17. February 2018 at 06:49

    Bob, MMT doesn’t say that debt doesn’t need to be serviced. MMT says that a currency issuing government should not need to issue debt in that currency in the first place. But if it has been issued, it needs to be serviced, just like any other obligation. Not servicing the bonds would punish the private sector that held bonds pretty severely.

  12. Gravatar of ssumner ssumner
    17. February 2018 at 07:51

    Doug, You said:

    “Is 2% as ceiling an average or a floor?”

    No, it’s an average, everyone agrees on that point.

    BC, Possible, but in the real world I don’t think that explains the difference. If it were, it would be 100% uncorrelated with liberal/conservative ideology.

    dtoh, I don’t know the answer to your first question. I do believe, however, that growth would be higher in the long run with a 3% NGDP target than a 10% NGDP target, due to second order effects like the taxation of capital income.

    Bob, Keep in mind that Tyler is basically saying to Noah “If you are a Keynesian and also believe there is growing monopolization then you should support stimulus right now.” I’m saying that’s not implication of the Keynesian view, or indeed any mainstream model. This has nothing to do with what I believe about the world or what Tyler believes, it’s about how Tyler characterizes the implications of Keynesian theory, which seems wrong to me.

    I can talk about the implications of Christianity, without myself necessarily being a Christian, or believing in their value system.

    As for MMT . . . SMH.

  13. Gravatar of Ralph Musgrave Ralph Musgrave
    17. February 2018 at 10:16

    “It’s incoherent to say, “I think demand side polices should always be stimulative.”” There is actually a sense in which that is a legitimate statement, as follows.

    Assume, to keep things simple, the stock of base money and national debt (private sector financial assets) is what the private sector wants to hold. If inflation is running at the 2% target level, then the real value of that stock will decline at 2%pa. Ergo to keep the value of the stock constant in real terms, government would need to run a deficit (a “stimulatory” policy)equal to 2% of that stock per year.

    Moreover, if the private sector’s desired stock of those assets remained constant relative to GDP (which it actually seems to in the very long term) and assuming there is real economic growth, then the deficit would need to be even larger.

  14. Gravatar of ssumner ssumner
    17. February 2018 at 14:12

    Ralph, All true, but that has nothing to do with demand stimulus. Stimulus doesn’t come from printing money, it comes from printing money faster than expected.

  15. Gravatar of John John
    17. February 2018 at 14:40

    As a member of Party X, I want an expansionary fiscal policy during times when Party X controls the government, but a contractionary fiscal policy when Party Y controls the government.

  16. Gravatar of Ralph Musgrave Ralph Musgrave
    17. February 2018 at 15:00

    Scott, Your point about “faster than expected” is true if Ricardian Equivalence is a powerful effect seems to me. Personally I think RE is near irrelevant: does the average household or small firm really sit down and work out what government is likely to do over the next two to four years and adjust their spending accordingly? I’ve never personally met anyone who does that.

    A more accurate model of the real world is where households and firms have some target amount of net liquid assets they are comfortable with, and adjust their spending accordingly, though even that is a big over-simplification of the real word: i.e. there a large amount of background noise.

    Far as I can see the empirical evidence does not give much support to RE, though I don’t claim to be an expert on that.

  17. Gravatar of dtoh dtoh
    17. February 2018 at 19:54


    You said, ” I don’t know the answer to your first question.”

    But this seems like a really important question because if you get it wrong, growth will be sub-optimal.

  18. Gravatar of Scott Freelander Scott Freelander
    18. February 2018 at 10:18

    Am I wrong in thinking Scott’s just making an obvious point here? I thought everyone knew monetary policy can only positively affect real GDP when there’s an output gap.

  19. Gravatar of Scott Sumner Scott Sumner
    18. February 2018 at 14:19

    Ralph, No, it has nothing to do with Ricardian equivalence. You are mixing up the issue of “Does X boost NGDP?” with “Does more NGDP boost RGDP?”

    dtoh, Fortunately, it doesn’t make much difference, as no one is promoting either 0% or 10% NGDP growth.

    If you’d asked me about 4% vs. 5%, I’d have answered that it makes virtually no difference.

    Scott, No, that’s not my claim. I think monetary policy can also affect RGDP at full employment. What I deny is that it can affect the trend rate of RGDP growth.

  20. Gravatar of dtoh dtoh
    18. February 2018 at 14:55

    Does 4 or 5 percent make no difference regardless of a) full potential RGDP growth rates, and/or b) whether or not the economy is running at full potential.

    I.e assuming potential (full employment) RGDP growth is 5%, but actual growth is 4%, is a 4.5% NGDP target optimal?

  21. Gravatar of dtoh dtoh
    18. February 2018 at 15:11

    Scott, or put another way….

    I don’t see how you fix sticky wages/prices if your NGDP target is below both potential RGDP and below actual RGDP growth.

    It seems to me you have to have some price component (above and beyond full employment RGDP growth) built into your NGDP target.

    Are you saying that the price component of the NGDP target can be anywhere from 0.001% to 2% and it won’t make any difference to RGDP growth?

    Or are you saying there doesn’t need to be a price component, which I think would be a complete refutation of the concept of sticky wages and prices.

  22. Gravatar of B Cole B Cole
    18. February 2018 at 16:11

    One interesting thing to watch for is that if the Fed creates a recession by over-tightening I think interest rates will hit zero on government bonds, as in Japan and Germany.

    At that point government borrowing becomes free, and perhaps the national debt can be paid down by issuing zero rate, perma-bonds.

    The world seems to have plenty of excess productive capacity, so I do not think the Fed should worry about inflation now.

    Another funny one: John Cochrane once ran a chart showing that at any time in the last 30 years the consensus of US macro economists was that inflation and interest-rates would head higher. Instead we have seen a near 40-year long secular decline in inflation and interest rates.

    I believe the risks and monetary policy are suffocation, not a mild bit of hyperventilation.

    Besides, a few years of full-tilt boogie boom times in Fat City would do wonders for the national soul.

  23. Gravatar of Scott Freelander Scott Freelander
    18. February 2018 at 20:05


    Thanks for the correction.

  24. Gravatar of Benjamin Cole Benjamin Cole
    18. February 2018 at 22:05

    OT but fun:

    Remember the soaring, searing double-digit inflation of the 1970s, that has frightened generations of macroeconomists?


    By the above PCE metric, inflation never got to double digits. It never even got to 9%! This metric was hatched by the freshwater (or more accuratley, dry gulch) Dallas Fed staffers, btw.

    BTW, the real economy expanded by 22% from Q2 1975 though Q1 1980.


    So yes, there was high single-digit inflation in the late 1970s. There was also very strong real growth.

    This also offers a way to “get out from under” our national debt.

    Run the economy hot, expand real output, generate lots of prosperity (unzome property for sure!) and run some moderate inflation for a decade.

    If you have a better idea how to handle a fresh $10 trillion in federal debt in the next 10 years, let me know.

    But start with the premise: We will have mountains of fresh debt soon.

  25. Gravatar of TheMoneyIllusion » Are hawks and doves simply confused? – Courtier en Bourse TheMoneyIllusion » Are hawks and doves simply confused? – Courtier en Bourse
    19. February 2018 at 14:41

    […] Le commentateur BC a tenté d'expliquer pourquoi il pourrait être rationnel pour certaines personnes d'être des faucons et que d'autres deviennent des colombes: […]

  26. Gravatar of ssumner ssumner
    19. February 2018 at 17:57

    dtoh, I’m saying inflation doesn’t matter, only NGDP growth matters. If there is a positive supply shock then RGDP growth will increase and inflation will decrease. That’s fine. That sort of disinflation does not hurt the economy.

    As far as specific numbers like 4% or 5%, either will lead to the economy staying close to the natural rate. Obviously I can’t say there is zero difference, but the difference would be very small, far less than 0.1%. Too small to notice.

  27. Gravatar of dtoh dtoh
    20. February 2018 at 02:04

    That sounds a little counter-intuitive. So just to be sure that I understand what you are saying. If we assume that if the natural rate is 3% (and there are no shocks) then, if we have a 4% NGDP target, we will get 3% real growth and 1% inflation, and if we have a 5% target we will get 3% real growth and 2% inflation (+/- a very small amount.)

    Also one more question. Under the same assumptions, if the NGDP target was 2%, would we still get 3% real growth but with negative 1% inflation?

  28. Gravatar of ssumner ssumner
    20. February 2018 at 11:01

    dtoh, Yes, Yes, that’s right, 3% either way. But all of these issues are much more complicated than you suggest, as it depends on where you start from. If you’ve been running 10% NGDP growth for many years, then suddenly dropping to 4% has a different impact than if the economy is already used to 4%. Ditto for suddenly dropping from 4% to 2%.

    Right now, the US has experienced almost 9 years of fairly steady 4% NGDP growth, so obviously that rate is no longer a “monetary shock”. We have adjusted to it–no reason to change it. But over the long run, 4% and 5% have almost identical effects, as the economy can equally well adjust to either. So I’d be equally happy with 4% or 5% as a long run policy.

  29. Gravatar of dtoh dtoh
    20. February 2018 at 18:03

    I get the complications, but on the other hand when you talk about the long run, I think you’re talking about the very, very long run. In the short run (at least for a few years and maybe longer), wouldn’t a higher NGDP target get you some extra real growth. And…. it’s not clear to me what would be the downside of doing that.

    Also wouldn’t a larger price component in your NGDP target give you more flexibility to deal with shocks (and to fix periods of incompetent Fed policy.)

  30. Gravatar of ssumner ssumner
    21. February 2018 at 09:53

    dtoh, The downside is destabilizing the economy, as we learned in the 1960s. The lesson of the 1960s is that we need a stable monetary policy, not an expansionary policy. After almost 9 years of stable growth in NGDP, why would we want to throw those hard earned gains in the trash? Have we forgotten the mistakes made in 1967?

Leave a Reply