Effective monetary policies are all alike, every ineffective policy is ineffective in its own way

More than 100 years ago Tolstoy got to the heart of what’s wrong with Keynesianism.  I don’t know why it took me so long to figure it out.

An effective monetary policy will steer NGDP growth at a fairly steady rate, such as the roughly 5% growth we experienced during the Great Moderation.  But suppose we don’t have effective monetary policy, what then?  Then we have Keynesian economics.

Keynesian economics purports to explain how all sorts of stuff will impact AD, which is roughly NGDP.  (NGDP and AD aren’t always defined as being identical, but in most standard Keynesian models than move in the same direction.)

Thus a change in business animal spirits, fiscal policy, consumer optimism, mortgage lending, the trade balance, etc, etc, will impact AD, according to the Keynesian model.  But if we have an effective monetary policy it will not affect AD.  I don’t think that’s very controversial.

Here’s where it gets controversial.  Many people will say; “but we don’t have an effective monetary policy, so the Keynesian model is valid.  The Fed won’t offset those real demand shocks; hence the shocks will end up moving AD.”

There are three problems with this Keynesian view.  First, as we saw during the Great Moderation, one cannot assume the Fed won’t offset the “non-monetary” demand shocks.  Quite often they did so, and very effectively.

But let’s say I’m wrong, and monetary policy is incompetent.  Even that doesn’t save the Keynesian model.  Because their next step is to assume “other things constant.”  They hold monetary policy constant when considering the impact of some expenditure shock in one of the economy’s sectors.  But you can’t hold monetary policy constant, because the entire concept is meaningless.  You can hold interest rates constant, or exchange rates constant, or M2 constant, or the monetary base constant, or TIPS spreads constant, or you can assume we constantly adhere to the Taylor Rule.  But there is no such thing as “monetary policy held constant.”  The term is hopelessly vague, and means different things to different people.

And of course even if you could hold monetary policy constant, any estimates of the impact of expenditure shocks would be worthless, because monetary policy in the real world would not be held constant, according to any criteria.  The Fed doesn’t hold interest rates constant.  It doesn’t hold M2 constant; it doesn’t hold the exchange rate constant.

There is only one effective monetary policy (stable NGDP growth), but there are a billion ineffective policies.  Policy can fail in a dizzying variety of ways.  And for each inept monetary policy, there is a completely different impact from a given expenditure shock.  That means there are a billion Keynesian models:

1.  There is the Keynesian model if Bernanke runs the Fed, and has three hawks on the FOMC.

2.  There is the Keynesian model if Volcker runs the Fed.

3.  There is the Keynesian model is G. William Miller runs the Fed.

4.  There is the Keynesian model if Bernanke runs the Fed, and there is one hawk on the FOMC.

and so on.

So when people talk about the effect of some “shock to AD, from a redistribution of income from low savers to high savers,” my response is “Huh?”  What are you talking about?  That isn’t science, it’s witchcraft.  How should I know how Bernanke would respond to that?  You convinced me he’s not following effective monetary policy.  So I’ve bought into that part of the Keynesian model.  But I have no idea which ineffective policy he is following.  And when I talk to others I realize that they are absurdly overconfident in their particular Keynesian model, mostly because they are blithely unaware of the problems I just laid out in this post.  And I’m calling out pretty much all the top macroeconomists in my field, including a certain unnamed Nobel Prize winner.

There is one Nobel prize winner I will name; Paul Krugman.  He does sort of understand this problem.  And his particular Keynesian model, his particular “ineffective monetary regime,” is nominal rates stuck as zero and no unconventional monetary stimulus.  Too bad that doesn’t describe our current Fed, or else Krugman’s version of the Keynesian model might have something useful to tell us about the impact of various expenditure shocks.

Just wait till we exit the liquidity tap, then you’ll see me really go ballistic in response to all the Keynesian drivel you see in the press.

PS:  Brad DeLong recently had this to say:

A Wicksellian is a believer that the key equation in macro is the flow-of-funds equation S = I + (G-T), savings S equals planned investment I plus government borrowing (G-T), and that the money market exists to feed the flow-of-funds an interest rate that has a (limited) influence on planned investment I. A Fisherian is a believer that the key equation in macro is the money market’s quantity theory equation PY = MV(i), and that the flow-of-funds exists to feed the quantity theory an interest rate that has a (limited) influence on velocity V.

Thus they have a hard time communicating. From the Fisherian viewpoint, the Wicksellians are talking nonsense because they spend their time on things that have a minor impact on velocity while ignoring the obvious shortage of money. From the Wicksellian viewpoint, the Fisherians are talking nonsense by ignoring the obvious fact that movements in money induce offsetting effects in velocity unless they somehow alter the savings-investment balance.

And it is we Hicksians, of course, synthesize both positions into a single unified and coherent whole…

I’m not sure DeLong is completely fair to Wicksell, but let’s say he is.  Then I’d say the Wicksell view is the view of people who are blithely unaware of the ideas in this post.  People who don’t even realize there is a “money reaction function problem.”  Some Fisherians underestimate the instability of velocity, but at least their framework is sound.  DeLong sounds like the sensible guy who combines the best of two extreme views.  I’d say that when you combine a sound theory with an unsound theory you end up with a half-baked model of the economy.

PPS.  These ideas were addressed from a slightly different perspective in this very early post, which appears to be Marcus Nunes’ favorite.

HT:  Thanks to John Papola for triggering this post with a thoughtful question.


Tags:

 
 
 

32 Responses to “Effective monetary policies are all alike, every ineffective policy is ineffective in its own way”

  1. Gravatar of marcus nunes marcus nunes
    19. December 2011 at 17:02

    Scott
    That´s right. That very early post was the “clincher” to me. The thing that made what came to be called MM (and NGDPT) make all the sense in the world.

  2. Gravatar of Ram Ram
    19. December 2011 at 17:28

    Hold the expected forward path for the Federal Funds Rate (FFR) constant. Hold the current setting of the FFR constant. A credibly temporary increase in the government’s budget deficit raises the natural rate of interest, stimulating aggregate demand. Simple as that. Why would the Fed hold these things constant in the face of a fiscal expansion? Perhaps fiscal stimulus is the easier sell in the current political environment. Is that so implausible?

    I don’t know why you feel you need to trash “Keynesians” as much as you do. Take a respected New Keynesian, like Woodford, and you’ll see him saying the optimal monetary policy framework is a flexible price level targeting rule (a fancier take on your preferred policy). I get that your favorite models don’t talk much about interest rates, inflation, etc. But lay down your arms, already. What do people like Woodford say about fiscal stimulus? It can work, but only if monetary policy lets it. Ideally the Fed would just do what’s necessary, but sometimes it doesn’t, and in those cases you can make fiscal policy work. Whatever happened to second-best problems in economics?

    You’d make a lot more friends and have a lot more influence, in my view, if you just said what is obviously the case from time to time: the New Keynesians and you basically see things the same way. You (like everyone else) have your minor differences with the benchmark New Keynesian models, but the worlds you’re describing are the same, and the first-best policies are the same.

    Then you can have an academic discussion trying to persuade New Keynesians to move towards your views, to focus more on the first-best rather than the second-best, etc. But as you’ve said yourself many times, what you’re writing about here isn’t academic. We’re talking about hugely important issues here, and the New Keynesians (who currently rule business cycle theory) see eye to eye with you on about 90% of how these things work. Karl Smith made this point well a little while back. Non-economists read posts like this and think economists are all over the place and have no idea what’s really going on, and what could do more to reduce the influence of yourself and the New Keynesians than that?

  3. Gravatar of Jason Jason
    19. December 2011 at 17:48

    Very interesting post, Scott; I had something of an a-ha moment while reading. This may be a bit strange.

    In physics, we have a the concept of “the vacuum of the theory” which represents a state where a) nothing interesting happens and b) more importantly, a theory of fluctuations from the vacuum is very simple.

    One of your claims is analogous to claiming that NGDP targeting represents the (true) vacuum. If we had a perfect NGDP target, there’d be no recessions or bubbles (the theory of fluctuations from the vacuum would be trivial).

    Another claim is that most other forms of monetary policy targets represent approximations to or imperfect models of the true vacuum so the actual monetary policy used will occasionally fail. The current model of the vacuum used by the Fed appears to be inflation targeting achieved by QE and a low interest rate target, but is actually unknown.

    Your billions of Keynesian theories represent the many possible expansions describing the fluctuations around various vacua.

    Frequently in physics we encounter a similar problem to another of your points: you have to specify the model of the vacuum you are expanding around. This would be analogous to “holding monetary policy constant” — specifying which monetary target your Keynesian theory represents an expansion around.

    To then boil down Sumner vs Krugman in this framework, we have 1) wanting the Fed to adopt an NGDP target for its vacuum theory vs 2) using a theory with interest rate targeting vacuum and IS-LM describing fluctuations at the zero lower bound. This latter theory suggests expansionary fiscal policy. The Sumner critique of it is that the Fed is not pursuing an interest rate target (the vacuum is not correct), so IS-LM is not valid. Fiscal policy won’t work because the current vacuum appears to be a 2% inflation target achieved by QE plus a low interest rate target, so the mechanism for fiscal policy won’t work in any Keynesian theory. The Krugman “critique” of the current situation is that neither fiscal expansion nor a different monetary vacuum is politically feasible so we are doomed.

    Summary:
    1. Sumner macroeconomic theory = NGDP target vacuum theory; no need to describe fluctuations
    2. Other macroeconomic theories = imperfect monetary vacuum theory + Keynesian model of fluctuations away from vacuum
    3. Most Keynesian models do not specify monetary policy vacuum so they can be tweaked (empirically fit) to work for any situation. They are theoretically useless.
    4. Some Keynesian models that do specify monetary policy vacuum use the incorrect vacuum.
    5. Our current vacuum appears to be a 2% inflation target with low interest rates and QE used to achieve it, but is vague enough to invalidate any Keynesian theory of the fluctuations.

    [I don’t know if I agree with 5 since we are somewhat near the true vacuum, the various monetary regimes should be good enough to do expansions around. IS-LM fluctuations appear to be invariant under different monetary policy vacua (IS-LM describes reality reasonably well under many common monetary policies without knowing what those policies are) — the reason it probably works is that most small fluctuations should be described by linear models and those with empirical coefficients represent the best guess we can make (an “effective field theory”).]

  4. Gravatar of Brito Brito
    19. December 2011 at 18:43

    “Thus a change in business animal spirits, fiscal policy, consumer optimism, mortgage lending, the trade balance, etc, etc, will impact AD, according to the Keynesian model. But if we have an effective monetary policy it will not affect AD. I don’t think that’s very controversial.”

    It’s a tautology, because you’re defining ‘effective monetary policy’ as something that ensures AD doesn’t drop, so saying that if the central bank ensures AD doesn’t drop then AD won’t drop is obvious.

    Also, I think the Keynesian model holds money (or the growth of such) constant (pick your measure, m2, m3 etc… the point is it holds constant any measure of money that would otherwise cause an AD boost should it expand). In an IS/LM model, this is evaluating a shift in IS without shifting LM. However I don’t think Keynesians ever actually assume LM doesn’t shift, UNLESS there is a liquidity trap where the curve cannot shift any further due to interest rates being close to zero. So really it seems the real problem is the concept of the liquidity trap, this other stuff seems irrelevant.

  5. Gravatar of Steve Steve
    19. December 2011 at 19:48

    Scott,

    You should include the original Tolstoy quote so that the unedumacated like myself don’t need to look it up.

    “All happy families are alike; each unhappy family is unhappy in its own way.”
    “• Leo Tolstoy, Anna Karenina

  6. Gravatar of Joe Joe
    19. December 2011 at 23:27

    This is the sort of thing that needs to be expanded into a longer formal essays, “A Treatise against Keynesian Economics.”

  7. Gravatar of Jack Rabuck Jack Rabuck
    20. December 2011 at 00:41

    This is very unrelated to this post, but my question formed while reading it, so I’m going to ask anyway.

    Is the effect of near zero percent interest rates a measurable deflationary pressure? What I mean is that if holding cash today has very little opportunity cost, will individuals and corporations not hold onto it during periods of stress or uncertainty, waiting for some other worldly signal that all is clear and there is no more turmoil? I always hear of easing as inflationary, and there is never much discussion of any countervailing pressure that may not make it so clearly cut.

    I’m a fourth year undergrad in Finance, so I fear my question will make me come off as rather unlettered. If there is a concept I should revise that would make this clear to me I would be happy to do the searching myself if you point me toward it.

  8. Gravatar of Adam Adam
    20. December 2011 at 05:01

    This doesn’t make much sense…

    If I have a banana and want to know what happens if I leave it on my counter for a month, you’re saying that’s a stupid programme of research because effective banana eating policy is to always eat the banana before it goes off. Except that sometimes I don’t…

    And then you say that it’s also a stupid programme of research because think how many DIFFERENT ways the banana could go off! It could be on table. Or in a fridge. Or in a dustpin. If you think about it there’s really nothing useful you can say even about bananas left on counters anyway. Except obviously there is…

  9. Gravatar of Ritwik Ritwik
    20. December 2011 at 05:47

    Scott

    I buy the basic point of ‘monetary policy held constant’ being meaningless. (To be fair, it just means interest rates held constant, but Market monetarists are saying something quite non-trivial here, so I grant the point).

    But proving the superiority of NGDP targeting over inflation targeting, or defining what an effective monetary policy is, are both somewhat truistic. In my understanding, the big questions are :

    1)Can the central bank ensure that it always hits its favourite target variable?

    2) If ensuring 1 above requires buying long dated private credit or real economy ‘stuff’, are we talking about monetary policy anymore? Do we basically just want a sovereign, fiat body to buy up loads of things under the guise of technocracy without any of the usual delays and debates of democracy.

    3) Most generically, should the target be a specific ‘level’ or a ‘corridor’. In the analogy of the Fed as the ‘captain’ of the economic ship, do we want ex-ante decisions of the path to be followed without regard to the storm or iceberg at hand?

    4) How would the recommendation change if we were to experience 5 years of 0% RGDP growth and 5% inflation?

    Somehow, you/ other market monetarists have managed to change the debate from these questions to ‘NGDPT>IT’ and ‘monetary policy is not interest rates’.

  10. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    20. December 2011 at 08:00

    Scott, I am a new reader to your blog, so please pardon me if my comment is redundant vs. previous posts. I am an economist gone bad: mortgage banker, investor, soldier etc. but have a longstanding respect for the monetarist viewpoint. I use a bastardized version of the quantity theory to guide my trading in interest rate futures, a 30 year avocation (which would not have endured if unprofitable). I know Friedman preferred M2 in quantifying money and you seem to be doing so as well. That never appealed to me on an intuitive level because it’s only M1 that can and is likely to be immediately spent. In addition, since the mid-1980s, would you not agree that it has been a better predictor of nominal GDP? I use my own modified definition of M1, which is the Fed’s version plus sweep accounts.

  11. Gravatar of ssumner ssumner
    20. December 2011 at 08:43

    Thanks Marcus.

    Ram, You said;

    “I don’t know why you feel you need to trash “Keynesians” as much as you do. Take a respected New Keynesian, like Woodford, and you’ll see him saying the optimal monetary policy framework is a flexible price level targeting rule (a fancier take on your preferred policy). I get that your favorite models don’t talk much about interest rates, inflation, etc. But lay down your arms, already. What do people like Woodford say about fiscal stimulus? It can work, but only if monetary policy lets it. Ideally the Fed would just do what’s necessary, but sometimes it doesn’t, and in those cases you can make fiscal policy work. Whatever happened to second-best problems in economics?”

    When I talk about Woodford I generally have positive things to say. This post isn’t directed at Woodford. I agree that he understands that monetary policy drives AD.

    I’m talking about the real world, which is full of Keynesians in and out of academia who have no idea what Woodford’s model says. They are 10 times more numerous.

    So I think you misunderstood the target of the post.

    I would add that I did not deny that non-monetary shocks have effects, indeed I said they do. I denied that there is a Keynesian “model” that explains these effects. The effect depends entirely on how the central bank reacts to the shocks. And the Fed almost never reacts to any sizable shock by holding constant the expected fed funds rate from now until infinity. That’s just not how they behave. We know that because sizable shocks impact interest rate futures.

    You said;

    “Non-economists read posts like this and think economists are all over the place and have no idea what’s really going on, and what could do more to reduce the influence of yourself and the New Keynesians than that?”

    Point well-taken. I accept that criticism.

    Maybe I’ll do a post responding to your point.

    Jason, That sounds right, but I’m no expert on physics.

    Brito, You said;

    “It’s a tautology, because you’re defining ‘effective monetary policy’ as something that ensures AD doesn’t drop, so saying that if the central bank ensures AD doesn’t drop then AD won’t drop is obvious.”

    Good point. Suppose I had said “effective inflation targeting policy.” I’d still make the same claim, and yet inflation is not equal to AD.

    Regarding “holding money supply constant,” you are outvoted 2 to 1 even in this comment section. And Ram is obviously a highly informed commenter. My point isn’t that it’s wrong to hold M constant, it’s that there is no obvious definition of “other things equal.” I prefer to hold NGDP futures prices constant.

    Sorry Steve.

    Thanks Joe.

    Jack, I agree, it’s probably not inflationary right now (although would be if rates were well above zero.)

    Adam, I’m saying it’s silly to model the future location of the banana with an abstract mathematical model that takes no account of the habits and preferences of the resident of the house.

    Ritwik, They can’t always hit their target, but that doesn’t matter. They can and should always hit their expected targeted. Expected NGDP growth should always be on target. That addresses the “storms” problem you raise–target the forecast!

    The amount and form of assets purchased depends on many factors:

    1. The size of the monetary base. This depends on both the trend rate of NGDP growth (policy goal) and the interest rate on reserves.

    2. The size of the national debt.

    In the US we can assume that a fairly high NGDP growth path, level targeting, would likely lead to nominal rates above 0%. In that case the base will be about a trillion dollars, much smaller than the $10 trillion in government debt held by the public. The Fed can also buy agency debt, foreign governement debt, but probably doesn’t need to. I’m not worried about them having to buy assets with high default risk.

    If they set the IOR above zero, things change. But why should they do that? If there is a problem of them having to buy risky assets, why not set the IOR below zero?

    In theory the issue you raise is valid, but in practice I think it’s highly unlikely to be a problem. The big demand for base money today is due to the low level of NGDP.

    Regarding point 4, the Fed should target NGDP and pay no attention to the P/Y split. If that outcome is bad, then time for Congress to improve the supply-side of the economy.

    Stephen, I’m afraid I just threw M2 out at random, I don’t think any monetary aggregate is informative. I focus on NGDP growth expectations, which I favor having the Fed target.

  12. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    20. December 2011 at 10:09

    Scott, I am surprised you don’t find any of the monetary aggregates informative. Here are the results of a Granger test with a two period lag between yoy percentage M1 growth vs. yoy nominal GDP percentage growth since 1947: M1 is significant as a possible cause of GDP at the 1% level while GDP as the cause of M1 is not significant at the 5% level. I used monthly data on M1 as defined above to include sweep accounts. GDP data are monthly by interpolation of quarterly data.
    At the practical level, this is extremely important because we know M1 with a one week lag while GDP lags many months. Precisely how do you propose to target nominal GDP if you don’t know what it is and can’t identify the appropriate parameters of leading metrics that are within the control of policymakers?

  13. Gravatar of Brito Brito
    20. December 2011 at 10:20

    Scott, you said:

    “Good point. Suppose I had said “effective inflation targeting policy.” I’d still make the same claim, and yet inflation is not equal to AD.”

    Scott, do you really think that AD can never drop if the central bank simply targets inflation?

    “Regarding “holding money supply constant,” you are outvoted 2 to 1 even in this comment section. And Ram is obviously a highly informed commenter. My point isn’t that it’s wrong to hold M constant, it’s that there is no obvious definition of “other things equal.” I prefer to hold NGDP futures prices constant.”

    But what is holding the LM curve stationary if not holding (a measure of) the money supply constant? It’s not interest rates, because a shift in IS will raise interest rates even if LM is held stationary. But perhaps we don’t have to define it, just define holding monetary policy constant as keeping the LM curve stationary, what’s wrong with that?

    Also, can’t Keynesians say exactly the same as you? Can’t Keynesians say that monetarist economics only applies as long as you hold fiscal policy constant? Can’t they also say that holding fiscal policy constant is hopelessly vague?

  14. Gravatar of ssumner ssumner
    20. December 2011 at 10:23

    Stephen, As you may know the best econometricians in the world basically abandoned monetary aggregates in the 1980s. Even many monetarists. There’s a reason for that, and your test doesn’t address their skepticism. The tests you discuss do not show causality, as in time series it’s easy to find one variable that leads another, without there being any marginal predictive power.

    Granger causality doesn’t show X causes Y, it shows X occurs before Y. That’s very different.

    I have posts discussing how the Fed should create NGDP futures contracts, and subsidize trading in that market. In the right side of this blog you can find links that connect to these posts.

  15. Gravatar of Charlie Charlie
    20. December 2011 at 10:30

    This is an interesting post, but requiring Keynesian economics to know the exact Fed response function is too high a burden. Even if there are a billion Keynesian models each mapping a different monetary response function, we could group them by characteristics and have reason to believe under certain conditions fiscal policy would have an effect. One of those characteristics would likely be the Fed “expecting to fail.” If Bernanke has an idea of an NGDP he wants to see, but is unable (politically) or unwilling to do what is necessary to hit that target. Even if your post gives me great pause about thinking I could estimate the fiscal multiplier, I’d still have reason to try fiscal policy.

    Second, this does undermine one of your big criticisms of fiscal policy, that the Fed will neutralize it. That assumes the Fed is following the specific version on ineffective policy that you ascribe to it, which is far from certain. In fact, this blog is one giant and persuasive argument that the Fed is running an ineffective monetary policy, so this post could be read as saying we should think really, really hard about the monetary policy response function, because figuring out the exact bad policy they are doing matters a lot.

    Lastly, it seems one could make this same post about NGDP targeting and some structural factors. We don’t know the exact mapping of investment/credit decisions or money demand to NGDP targeting. We don’t know how much real growth vs. nominal growth various targets (3%, 5%, 7% for two years then 5%?) will create. We don’t know exactly how the market/economy will respond to the Fed switching regimes and the million different ways they might attempt to do it. I don’t think that undermines NGDP targeting as the best policy we could choose. Just because we don’t know everything, doesn’t mean we don’t know some things.

  16. Gravatar of ssumner ssumner
    20. December 2011 at 10:32

    Brito, You said;

    “Scott, do you really think that AD can never drop if the central bank simply targets inflation?”

    No, in order to target inflation one must adjust AD to offset AS shocks, so I’m not saying that. I’m not even saying one of the Keynesian demand shocks couldn’t momentarily throw you off your inflation target. I’m say that if we are going to model these “demand shocks” we need to model how far they throw monetary policy off course.

    Regarding holding M constant, sure you can do that–maybe the LM model says you should do that. But you can also hold lots of other things constant. The question is which thing is it most USEFUL to hold constant. I say you’d want to hold constant the actual real world monetary reaction function, and we don’t have a clue as to what that is, hence we don’t have a clue as to how all these Keynesian “demand shocks” impact NGDP.

    You said;

    “Also, can’t Keynesians say exactly the same as you? Can’t Keynesians say that monetarist economics only applies as long as you hold fiscal policy constant? Can’t they also say that holding fiscal policy constant is hopelessly vague?”

    For lots or reasons it’s not at all symmetrical, especially during “normal times” when rates aren’t zero. Then the Fed has to only slightly adjust monetary policy, in order to offset non-monetary demand shocks. Only a slight adjustment in the monetary base.

    Second, monetary policy is essentially costless, and moves last. Fiscal policy is costly, and moves first. So as a practical matter it is monetary policy offsetting fiscal stimulus, not the reverse.

  17. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    20. December 2011 at 10:39

    Thanks for your quick reply, Scott. Yes, I am well aware of the post hoc fallacy and the limitations of statistical inference. As I said, however, I have been trading interest rate futures for over 30 years and would not have continued if it weren’t profitable. Money leading GDP is the critical information I need to realize a profit because it is indicative of whether interest rates are currently too high or too low to generate the nominal GDP that will be consistent with one or the other of the Fed’s mandates. (It is really only one of the two that they are targeting at any given time.)
    Let’s try a real-time test that will take up to one year to complete: YOY M1 at 13% in the past 12 months will yield a sharp acceleration in nominal GDP growth in 2012. Money market rates (I trade Eurodollar futures) will rise and, in fact, have already begun to do so. I don’t even care how much they rise; just getting the direction right is all I want.
    No need to reply. Your blog is a wonderful addition to the academic and political discussion; I enjoy it immensely and wish you a very merry Christmas.

  18. Gravatar of ssumner ssumner
    20. December 2011 at 10:41

    Charlie, You said;

    “One of those characteristics would likely be the Fed “expecting to fail.” If Bernanke has an idea of an NGDP he wants to see, but is unable (politically) or unwilling to do what is necessary to hit that target. Even if your post gives me great pause about thinking I could estimate the fiscal multiplier, I’d still have reason to try fiscal policy.”

    You’d think that was true, but it probably isn’t. Failure is relative, it’s likely that the Fed has simply set the bar lower, because they find QE distasteful. Suppose their inflation floor is now 1% instead of 2%, and they do QE when inflation falls below that level. Then if the fiscal policymakers do more stimulus, the Fed simply does less. Only if the fiscal policymaker boosts inflation high enough where no QE is needed, where we are in the world of positive interest rates, does it really help.

    Now astute critics like Andy Harless have shown my argument is a bit too simplistic, but I’m still not convinced it’s all that far off course.

    You said;

    “Second, this does undermine one of your big criticisms of fiscal policy, that the Fed will neutralize it. That assumes the Fed is following the specific version on ineffective policy that you ascribe to it, which is far from certain. In fact, this blog is one giant and persuasive argument that the Fed is running an ineffective monetary policy, so this post could be read as saying we should think really, really hard about the monetary policy response function, because figuring out the exact bad policy they are doing matters a lot.”

    Don’t get carried away, as this is something I’ve acknowledged dozens of times–but see my reply to your first point. I’ve never denied that fiscal stimulus might have boosted NGDP.

    Regarding your third point, I favor targeting the forecast. That’s the best any policy can do. You can do that with monetary policy if you choose (we choose not to.) I can’t even imagine trying to do that with fiscal policy.

  19. Gravatar of ssumner ssumner
    20. December 2011 at 10:42

    Stephen, There’s no way I can argue the Ms with you, as it’s not my field. But didn’t they fail to predict the severe 2008-09 slump?

  20. Gravatar of Brito Brito
    20. December 2011 at 10:45

    Scott, you said:

    “Regarding holding M constant, sure you can do that-maybe the LM model says you should do that. But you can also hold lots of other things constant. The question is which thing is it most USEFUL to hold constant. I say you’d want to hold constant the actual real world monetary reaction function, and we don’t have a clue as to what that is, hence we don’t have a clue as to how all these Keynesian “demand shocks” impact NGDP.”

    I don’t think we have NO idea, firstly it depends on the bank. I think the Fed has a Taylor type reaction function but with a low coefficient on output gap fluctuations and a very high coefficient on inflation fluctuations. Didn’t Taylor himself actually work out the reaction function using data in the 80s and 90s?

    “For lots or reasons it’s not at all symmetrical, especially during “normal times” when rates aren’t zero. Then the Fed has to only slightly adjust monetary policy, in order to offset non-monetary demand shocks. Only a slight adjustment in the monetary base.

    Second, monetary policy is essentially costless, and moves last. Fiscal policy is costly, and moves first. So as a practical matter it is monetary policy offsetting fiscal stimulus, not the reverse.”

    But in a liquidity trap it would appear that the central bank needs to engage in very bold and committed actions, not slight adjustments. So there is co-dependence, when rates are near zero the effectiveness of fiscal policy depends on the boldness of monetary policy, and the effect of unconventional monetary stimulus depends on how bold fiscal authorities are being. Furthermore, what about a hybrid policy? That is, fiscal expenditures financed by money printing? This has to be analysed using Keynesian economics, because it is an increase in government spending.

  21. Gravatar of Matt Waters Matt Waters
    20. December 2011 at 10:56

    “But in a liquidity trap it would appear that the central bank needs to engage in very bold and committed actions, not slight adjustments. So there is co-dependence, when rates are near zero the effectiveness of fiscal policy depends on the boldness of monetary policy, and the effect of unconventional monetary stimulus depends on how bold fiscal authorities are being. Furthermore, what about a hybrid policy? That is, fiscal expenditures financed by money printing? This has to be analysed using Keynesian economics, because it is an increase in government spending.”

    There’s a simple solution here. Either work in a liquidity trap to the extent they’re used.

    Even without printing money, fiscal stimulus increases AD by increasing velocity. Treasuries are considered a cash equivalent and the government mobilizes money that’s otherwise idle.

    The government printing money has a less direct effect, which is why Krugman doesn’t like it, but buying longer-term bonds does increase velocity by displacing current long-term bond investors and making them choose between cash or actually investing or spending in something.

    The real issue is political. If we only engage in fiscal/monetary stimulus to the extent where we prevent the economy from getting worse, then that’s all we’ll get. That’s why Japan has been in a holding pattern. But if the politics change to realize we should do one or both until we get back to full employment, then that’s what we’ll get. Doesn’t matter which one we choose. If we do choose fiscal stimulus, that just offsets the monetary stimulus that would have been politically feasible otherwise.

  22. Gravatar of Effective monetary policies are all alike, every ineffective policy is ineffective in its own way « Economics Info Effective monetary policies are all alike, every ineffective policy is ineffective in its own way « Economics Info
    20. December 2011 at 11:00

    […] Source […]

  23. Gravatar of Brito Brito
    20. December 2011 at 11:03

    “The government printing money has a less direct effect, which is why Krugman doesn’t like it, but buying longer-term bonds does increase velocity by displacing current long-term bond investors and making them choose between cash or actually investing or spending in something.”

    Right, the portfolio adjustment effect, but the thing is there are many places investors can put their money, much of it will be international, and if they’re large banks it will be at reserves at the central bank that pay interest, I’m unconvinced that buying long term government bonds will have that much of an effect itself.

  24. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    20. December 2011 at 11:05

    Scott, M1 has worked for me, but I must stipulate that it is only one of several variables I use. To document my position I submit my comments on “The Wall Street Journal” online in response to two relevant articles. The first was in August 2007:
    http://blogs.wsj.com/economics/2007/08/14/consumer-price-data-could-ease-fed-concerns/
    The second was in anticipation of the economic recovery in early 2009:
    http://blogs.wsj.com/economics/2009/01/28/money-supply-reverses-course-as-an-indicator/
    I suppose the third example is my above and yet to be borne out claim that GDP will accelerate in 2012 but–of course–I may just be the personification of post hoc.

  25. Gravatar of ssumner ssumner
    21. December 2011 at 07:24

    Brito, Under any sort of plausible Taylor Rule, the fiscal multiplier is zero, isn’t it?

    As far as your second point, sure, you can monetize the debt. But if the central bank is that bold, why would you need fiscal stimulus in the first place? If we’ve got a problem, it’s probably because the central bank is not willing to aid fiscal policy.

    Matt, You said;

    “Even without printing money, fiscal stimulus increases AD by increasing velocity. Treasuries are considered a cash equivalent and the government mobilizes money that’s otherwise idle.”

    People simply aren’t paying attention to what I’m saying, and it’s starting to drive me crazy. Sure, fiscal stimulus can increase V, but that has no effect on M*V, which is what we care about. The Fed certainly isn’t going to hold M constant, indeed they’ve nearly tripled the base in recent years! The Fed controls M*V, not Congress.

    Stephen, You haven’t answered my question. What were the growth rates of M1 during 2007 and 2008?

    It’s fine to provide links, but I only have time to read a small portion of them. So you also need the answer provided in your comment.

  26. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    21. December 2011 at 09:33

    Scott, I didn’t understand that you were asking for supporting data. The links I provided merely document that the undisclosed model I use has put me on the right side of the resulting economic environment. M1 (as defined above to include sweep accounts) is one of the variables in my specification and I think you will agree that the growth rates in this metric over the past five years have been positively correlated with second derivatives in nominal GDP as well as job growth:
    2006– 1.1%
    2007– 1.6%
    2008– 10.5%
    2009– 5.9%
    2010– 6%
    2011– (12 months to November) 13.1%
    At a minimum, I think you will agree that the M1 data support your position that the recession was caused by overly tight Fed policy, but you may not share my opinion that current policy is far too lax and will generate a too-rapid recovery in 2012.

  27. Gravatar of Brito Brito
    21. December 2011 at 11:20

    Scott, you said:

    “Brito, Under any sort of plausible Taylor Rule, the fiscal multiplier is zero, isn’t it?”

    Only if the Taylor Rule keeps output at or above the natural level. I can’t find my lecture notes on this (I’m postgrad student), but I’m pretty sure that the rule Taylor estimated the Fed was operating under at the time put too heavy a coefficient on an inflation shock, and too small a coefficient on the output gap shock, which would mean that the central bank reaction were there to be a huge shock to output but not as big a shock to inflation/deflation may not be adequate to restore output to its natural level quickly, leaving a role for fiscal stimulus.

    “As far as your second point, sure, you can monetize the debt. But if the central bank is that bold, why would you need fiscal stimulus in the first place? If we’ve got a problem, it’s probably because the central bank is not willing to aid fiscal policy.”

    I think debt monetization, if it prevents governments from cutting public finances big time, is a form of fiscal stimulus; it reassures the public that spending wont fall, or in other words increases their expectations of public spending, so I would happily call that fiscal stimulus.

  28. Gravatar of Shane Shane
    21. December 2011 at 20:15

    Our current ineffective monetary policy regime is quite easy to define, actually:

    1. Fed Funds rate at zero for indefinite future (i.e., until unemployment falls dramatically)

    2. QE when deflation threatens to rear its head.

    3. Bernanke explicitly asks for more fiscal stimulus.

    4. Bernanke thereby implies that he will allow catch-up growth, but that he won’t create it.

    Under this, our current (and very well defined) monetary policy regime, fiscal stimulus is indeed effective.

  29. Gravatar of ssumner ssumner
    22. December 2011 at 07:23

    Stephen, That data seems to blow your argument out of the water. Double digit M1 growth in 2008 and the economy falls off the cliff? You must see something I missed.

    Brito, But if they put a 100% weight on inflation the fiscal multiplier is still zero–that’s the standard view.

    I don’t follow your second point. Any open market purchase is “debt monetization.” Unless they pay interest on reserves, or course.

    Shane, Even if we buy your assumptions, your conclusion is wrong. The assumptions imply that if the fiscal authorities do more, we’ll get less QE. Tell me where my logic is wrong? Perhaps if they do so much fiscal stimulus that no QE is needed it might create a faster recovery. But seriously, how likely is that?

  30. Gravatar of Stephen Delos Wilson Stephen Delos Wilson
    22. December 2011 at 08:19

    Quite the contrary, Scott, the data are consistent with my claim that money growth leads nominal GDP growth. The lag I’ve found to give the best fit is YOY lagged 18 months. We had a reversal of nominal GDP in 2009: from the trough in the second quarter of -3.9% nominal YOY GDP growth reversed to a peak rate of 4.9% in the third quarter of 2010. The subsequent decline in YOY nominal GDP growth is consistent with the subsequent (2009 and 2010) decline in the rate of M1 growth. The coming year will see the effects of the far more rapid M1 growth in 2011.

  31. Gravatar of Shane Shane
    22. December 2011 at 10:51

    The assumptions as written above imply that the fed will withhold only the amount of stimulus that would have been needed to prevent deflation.

    Based on actual practice though, it seems doubtful that the Fed would actually even do this. QE 1 started right at the height of the stimulus. Maybe the Fed would have announced a higher dollar figure for QE in its absence, but as you always say, it’s not the dollar amount per se that makes QE effective so much as that it signals that the Fed wants to change expectations about growth.

    So unless not passing the stimulus would have pushed the Fed toward some sort of level targeting–a doubtful conclusion–then it’s very likely the stimulus was effective.

  32. Gravatar of ssumner ssumner
    23. December 2011 at 07:35

    Stephen, Monetary policy affects the economy with almost no lag at all. In addition, that 18 month lag you claim to find, (which I don’t see) would completely fail during other periods. It’s not a stable relationship.

    The monetary shocks that are easiest to identify occurred during the interwar years, and they all affected the economy within a month or two.

    Shane, I am almost certain Bernanke would not have allowed a depression, that’s simply not his background–he’s a Depression scholar who feels passionately that the Fed was too passive in the 1930s. I don’t know if he would have done level targeting, but surely the Fed would have done something much more aggressive w/o fiscal stimulus. We might well have gotten even faster NGDP growth.

    I do understand why not everyone agrees with me about Bernanke, most people haven’t studied his writings as much as I have, and I’m also making a psychological judgment. But NO ONE is giving me any good reasons why I should not hold these views.

Leave a Reply