Two mules pulling in opposite directions: Why independent central banks fail

During the 1990s the concept of an “independent central bank” became much more fashionable.  We were all trying to insulate central banks from political pressure, so that we could achieve German-style inflation.  In one sense it worked.  The Fed, BOJ and ECB have delivered German-style inflation.  But nonetheless the experiment has been a failure.  It failed in Japan, and is now failing in America.  It’s useful to consider why.

The Achilles heel of independent central banks is that the fiscal authority believes, rightly or wrongly, that fiscal policy also affects the inflation rate.  Thus when you have a disagreement between the fiscal and monetary authorities as to the appropriate rate of inflation, the two will end up pulling in opposite directions, producing massive economic waste in the form of higher than necessary marginal tax rates to pay for wasteful deficit spending.

We all saw this (two mules) problem in Japan during the 1990s and early 2000s, but somehow assumed (or at least I did, Krugman didn’t) that our policymakers were more sensible.  I also think most economists missed this problem because they don’t think of the fiscal authority as influencing the inflation rate.  For some reason even those economists who look at stabilization policy in the old-fashioned two-pronged way (use of both fiscal and monetary stimulus) tend to think that monetary policy affects inflation whereas fiscal stimulus affects real growth.  I have talked about this issue quite a bit, but still haven’t heard any good explanations for the ubiquity of this strange way of framing stabilization policy.  After all, both fiscal and monetary stimulus impact AD, and hence both influence output and prices.

So we go into a severe recession.  The central bank decides “inflation is not a problem” and thus keeps monetary policy unchanged.  The fiscal authority decides “real growth is too low” and thus adopts fiscal stimulus.  Yet few people see the inconsistency here.  I’ll bet you could find plenty of Congressman in Washington who support fiscal stimulus, and who favor more aggregate demand, but would oppose any Fed attempts to raise inflation.  It makes no sense, but that’s the crazy world we live in.  The mules in Congress don’t even know that someone else in pulling in the opposite direction.

My suggestion is that we end the independence of central banks, but replace it with something else–an explicit, legal, central bank target.  Let’s suppose the Congress instructed the Fed to target 2% inflation.  Now assume Congress wants more growth because we are in a recession, but because inflation is currently 2% the Fed doesn’t want to ease.  The fiscal authorities could instruct the central bank to aim for 2% inflation in the long run, but allow a bit more inflation during the current recession, at the expense of a bit less that 2% inflation during the subsequent years.  In other words the Fed would still have some discretion, even though their long-run mandate would be 2% inflation.  They would have the legal authority to tell the fiscal authorities “OK, we’ll provide a bit more inflation right now, as long as you understand that it is being “borrowed” from future inflation.  We intend to run below 2% inflation during the next boom.”  This would allow for some fruitful policy coordination, while still protecting the central bank from pressure to alter its long run inflation target.

You might be thinking; “Lower than average inflation during booms and higher than average inflation during recessions.  What a great idea for macroeconomic stabilization.  How can we formulate that into a simple and easy to understand nominal target?”

Seriously, Congress needs to decide on some sort of explicit policy goal for aggregate demand.  Whether it be inflation, the price level, or NGDP, the point is to have clearly spelled out goals so that they and the Fed are pulling in the same direction.  Indeed if the goals are spelled out, there is no reason why Congress should ever have to do any pulling.

And that brings me to the final point.  Even if this doesn’t improve monetary policy at all, at least I will be spared from annoying arguments with Keynesians who say; “Well yes, monetary policy could work in theory, but since the Fed is being so conservative we have to rely on fiscal stimulus.  With the Fed no longer independent, there would no longer be any excuses for policy failures.  The voters already blame Bush and Obama for the high unemployment; wouldn’t it be nice if the voters were correct?



31 Responses to “Two mules pulling in opposite directions: Why independent central banks fail”

  1. Gravatar of Doc Merlin Doc Merlin
    20. June 2010 at 13:29

    Oh the other hand, a non-independent fed means that the fiscal authority can pressure the fed to pay for their profligate spending… thus leading to too much, not too little monetary expansion.

    How about instead we make government money based on some sort of commodity? And private money is allowed to issue its own money denominated in any way they wish? We get the benefits of a independent monetary authority, and we get the benefits of a nonindependent monetary authority. It is win-win.

  2. Gravatar of scott sumner scott sumner
    20. June 2010 at 13:32

    Doc Merlin, Not under my proposal. I favor replacing the independent central bank with an explicit monetary policy target. Of course you could argue that Congress could pass a law changing the explicit target, but it’s equally true that they can pass a law revoking the Fed’s independence.

    We tried commodity money with the gold standard; it didn’t work.

  3. Gravatar of Doc Merlin Doc Merlin
    20. June 2010 at 14:03

    Commodity government money only work if you have the ability of the private market to issue private non-commodity money.
    This is evidenced by commodity money working fine for over 100 years, then failing horribly after we removed legal private money issuance.

  4. Gravatar of david david
    20. June 2010 at 15:18

    Scott, I think your indefatigable patience to walk arguments through with your commenters has encouraged people to march in and make extremely curious statements on economic issues without feeling obliged to defend said curiousness. :P

    Doc Merlin, how do intend to counter the procyclicity of commodity monies? Or overcome the real transaction costs incurred when many currencies exist within an economic trade region? What happens when the private provider prohibits the sale of private money to third parties (e.g., as company scrip often did) – is the state entitled to refuse to enforce this prohibition?

  5. Gravatar of John John
    20. June 2010 at 17:54

    @doc merlin
    “This is evidenced by commodity money working fine for over 100 years”…

    No. It did not work fine. Actually, it was pretty terrible. I don’t really feel any more need to defend this statement than you do to defend yours, although I am very willing to assert vigorously that it is correct.

  6. Gravatar of Christopjer Hylarides Christopjer Hylarides
    20. June 2010 at 19:26

    Doc Merlin is taking the same romantic view of the 19th century that so many gold standard fans take. They ignore the banking panics as well as massive amounts of inflation that happened when new gold was discovered in California, the Yukon, South Africa, and Australia.

    There was also always a massive amount of pressure on the classical gold standard. Farmers hated it, they wanted a silver standard. Manufacturers only liked it if they were trading with europe. In south america and asia, silver ruled. Government of course liked fiat money, espexially when it was easy/acceptable to inflate in the early years of it.

    While I’m a big fan of free banking and private currency (they’re some of the reasons my native Canada didn’t experience deflation or any bank failures during the depression), it wasn’t all sugar and spice.

  7. Gravatar of Larry Larry
    20. June 2010 at 20:15


    Is this as shocking as I think?

    “With a goal of achieving a 3% nominal growth rate and ending deflation, Prime Minister Naoto Kan’s aides identify in the blueprint scores of “national strategy” projects in key areas like the environment, health care and infrastructure exports to Asia.”

  8. Gravatar of Benjamin Cole Benjamin Cole
    20. June 2010 at 20:47

    First time reader here. I am glad to see someone puzzling over monetary policy, and not in the usual “I am a conservative therefore I believe in a tight money supply” way.
    I have been watching MZM dropping, and yet you have commentators talking about an “accommodative Fed policy.”
    They blab about gold, and forget it is Chinese buying gold with newfound disposable income.
    And that the global money supply might be more important than the US money supply, if we have free and open borders for the movement of capital (we do, mostly).
    I look forward to doing some reading of your earlier posts, but I see a lot that I like in looking through.

  9. Gravatar of Doc Merlin Doc Merlin
    20. June 2010 at 21:30

    “Doc Merlin is taking the same romantic view of the 19th century that so many gold standard fans take. They ignore the banking panics as well as massive amounts of inflation that happened when new gold was discovered in California, the Yukon, South Africa, and Australia.”

    Yes, inflation happened, but problems were temporary and we were able to scale back from them. Yes, panics happened, I never said that the gold standard was perfect, or that we should return to it. I said that if the government wishes to issue money, some sort of commodity standard is preferable to a captured monopolistic central bank. In other threads I have stated that I like the idea of a fed that has a hard target (compared to the one we have now), but I wanted to point out that commodity money + free banking has the same benefits that a captured fed has, while still allowing for some market discretion not found in a captured, monopolistic fed.

    “While I’m a big fan of free banking and private currency (they’re some of the reasons my native Canada didn’t experience deflation or any bank failures during the depression), it wasn’t all sugar and spice.”

    Well, of course not, but properly managed it makes it hard for government to borrow in times of peace, so we didn’t have these huge peacetime deficits, and their accompanying problems. It also added competition and flexibility into the system, so everyone didn’t have all their eggs in one basket, as they do in a pure commodity standard or a pure monopoly fiat money issuer.

    I’ve been commenting here for a long time, and yes, Scott is one of the best bloggers imo, precisely because he walks people through ideas and issues. He’s been able to raise his blogging status and convert quite a few people to his side, because of his clearness and patience.

    Anyway, to answer your questions one at a time.
    “Doc Merlin, how do intend to counter the procyclicity of commodity monies?”

    That is countered by the flexibility of private money. The supply of private money grows/shrinks automatically in response to market forces, which makes it much more forward looking than fiat money. This is why Scott wants prediction market for targeting monetary policy. Well, if we have a free market in money, it doesn’t need a prediction market, because it does it itself.

    “Or overcome the real transaction costs incurred when many currencies exist within an economic trade region?”

    Because of computers this is actually the easiest one to answer. Most transactions nowadays are electronic (credit cards, debit cards, ATM cards, electronic checks etc). It is possible and easy to minimize transactions costs as a result. Also, as Krugman has pointed out (in a rare case where I agree with him), there are also real costs to having a single money within a diverse geographic area.
    Mexico is an interesting example, because people rightly don’t trust Mexican monetary authorities. As a result, savings accounts can be denominated in silver or in Mexican currency. This doesn’t result in too high transaction costs, because, in the modern world, currency transaction costs between banks are really, really tiny.

    “What happens when the private provider prohibits the sale of private money to third parties (e.g., as company scrip often did) – is the state entitled to refuse to enforce this prohibition?’

    The state does refuse to enforce this same prohibition now for some other things, so I don’t see the problem.

  10. Gravatar of MMJ MMJ
    21. June 2010 at 03:18


    1) I was under the impression that New Keynesians favour monetary over fiscal policy, and that Krugman, DeLong, Mankiw et. al. are in that camp..?

    2) You seem to be suggesting something qualitatively similar to Bernanke’s price level + trend target outlined in 2003. Is that the correct interpretation?

    Thanks and regards, MMJ.

  11. Gravatar of W le B W le B
    21. June 2010 at 04:20

    An elected politician with a mandate from and a responsibility to the public is a fool who in the last two years of the present climate did not want to seize back the headline direction of monetary policy from the central bank. The determination of the supply of money is a political choice. It either lies with the elected or the unelected.
    You are right the other day to see class as an active ingredient in the UK mix. The old class system has imprinted itself on the new meritocracy that now makes up the leadership of the political class in all three parties. The new class with these inherited authoritarian instincts has assumed and actually further developed the state’s centralisation.
    The former land and trade divide has persisted as the divide between a professional and academically influenced political class and the get on and do it ‘university of life’ business class. Business here understands trade but has a balance sheet view of economics – micro as macro – whilst the politicians don’t understand business, though from the safety of the public sector they think they do.
    When businesses small, medium and large were running for cover in 2007 instinctively sensing a downturn, central bankers were raising rates and politicians were campaigning for ‘tougher action’. They just hadn’t picked up the signals that business was getting and acting on.
    But the real tragedy is that the belief in the relationship between MV and PT was demonized in the Seventies and Eighties and is lost to this political class. Hence their concentration on the fiscal situation and the assumption that monetary policy is about interest rates rather than about the quantity of money and the velocity of its circulation.

  12. Gravatar of Doc Merlin Doc Merlin
    21. June 2010 at 04:33

    Awesome, Russ’s interview of you is up.

  13. Gravatar of scott sumner scott sumner
    21. June 2010 at 07:10

    Doc Merlin, I agree that the abandonment of private currency issue made the gold standard less stable, but still worry that the value of gold may be too unstable for a modern economy. A deflation of 3% today is vastly more harmful than in 1840.

    David, I agree that there are some potential problems with Doc’s proposal that make it too risky to adopt. I share your fear of unstable gold values. But he may be right about it being fairly effective while it was in operation.

    Christopher, One small correction–Canada did experience deflation in the early 1930s. But you are right that they had a much better banking system than we did (and still do.)

    Larry, Thanks, I may do a post on that. (Notice they don’t mention monetary policy, however.)

    Benjamin, Thanks. The FAQs link gives an overview of my ideas.

    Doc Merlin, Those are good points. I think George Selgin’s arguments for free banking are the strongest. He contemplates a stable monetary base. However because the stock of gold grows at a very steady rate, the arguments also apply fairly well to free banking combined with a gold standard.

    MMJ, What new Keynesians favor is a complicated question. Bernanke is clearly a new Keynesian, but in 2008 advocated fiscal expansion, a very non-new Keynesian policy. So I think they are all over the map. But yes, in theory they prefer monetary policy.

    I have posted on Bernanke’s 2003 paper, and criticized him for not adopting the policies that he recommended the Japanese adopt when rates hit zero. I do support level targeting (of prices, or better yet NGDP.)

    W le B, I agree that the political class is pretty clueless about monetary policy. And I was aware of the drift in Britain toward a sort of liberal authoritarianism (if that’s not an oxymoron) with even Labour supporting all sorts of restrictions on freedom for paternalistic reasons.

    Doc Merlin, Thanks, I have a new post that links to it.

  14. Gravatar of Christopher Hylarides Christopher Hylarides
    21. June 2010 at 07:42

    @Doc Merlin,
    Bretton Woods was still a (bad) gold standard and it didn’t stop the USA from running large deficits as well as inflating its currency from both war and the great society.

    A government commodity standard doesn’t stop the government from inflating anyways, to say nothing of more transparent actual revaluations. In theory it would, but in practice it almost never did. Even countries that followed the rules better than others, like Britain, got screwed due to the way the American Fed locked away gold instead of issuing more currency.

    @scott sumner
    Yes, what I meant was monetary deflation (it’s always easy to forget to specify that!), of which there was very little in Canada. The deflation in Canada wasn’t monetary, but price as demand tapered off due to unemployment due to less exports to the United States. At least that’s what I’ve been told is one of the reasons Canada experience no bank failures.

  15. Gravatar of Stephen Morris Stephen Morris
    21. June 2010 at 14:46

    As an interested lay person with no special expertise, I have often wondered why interest rates aren’t set according to some public adaptive monetary policy algorithm which would allow the market (in effect) to set the short term rate, thereby allowing fiscal policy to be used as the single policy lever.

    For example, targeting the yield curve as follows:

    I = (LTBR – a) + b*(f – t)


    “I” is the target Funds rate

    LTBR is the long term bond rate (e.g. the ten year Treasury)

    “f” is the most recent measure of inflation according to the index adopted by the Fed

    “t” is the Fed’s (published) target for this rate

    “a” is a published constant embodying the yield curve spread at equilibrium inflation

    “b” is a published coefficient

    If one assumes that at equilibrium inflation (f = t), the yield curve should have a spread of, say, 2%, then the constant “a” would have that value.

    The coefficient “b” sets the sensitivity of the adaptive system. If it is assumed that the maximum deviation above target inflation should be 1%, and that the yield curve should then be flat, “b” would have a value equal to “a”. Thus, for example:

    I = (LTBR – 2%) + 2*(f – 1%)

    Obviously, as with any auto-pilot system, it would require an override in an event of force majeure, but would it be any worse than having fallible human beings deciding the matter?? And everyone would know what the policy was going to be . . . years in advance.

  16. Gravatar of scott sumner scott sumner
    22. June 2010 at 06:04

    Christopher, I hate to say this, but I recall that even the monetary base fell in Canada. The key difference is that they didn’t have bank failures, as you indicated.

    Stephen, That’s similar to the Taylor Rule. I favor directly targeting the price level or NGDP, as interest rates are not always a reliable indicator of monetary policy.

  17. Gravatar of Mike Sandifer Mike Sandifer
    22. June 2010 at 07:35

    Scott, a question just occurred to me today.

    Can central banks simply buy their own governments’ bonds, retire the debt, and then raise bank reserve requirments to counter inflation, even at full economic capacity(maybe an awkward way of saying sans slack in the economy)?

    If this is a stupid question, then I apologize, but it certainly wouldn’t be out of character.

    If it would work this way, then deficits and debt really don’t matter.

  18. Gravatar of Mike Sandifer Mike Sandifer
    22. June 2010 at 08:01

    It occurs to me that there would be forex consequences for the above, but what if every central bank did this?

  19. Gravatar of Christopher Hylarides Christopher Hylarides
    22. June 2010 at 15:49

    But then why didn’t banks fail in Canada? There was still a fractional reserve system there.

  20. Gravatar of pacer pacer
    22. June 2010 at 18:27

    Just give me a system that allows my money to hold its purchasing power in a coffee can. The biggest flaw with the Fed is that its idea of ‘price stability’ is 2% inflation as measured by the government’s screwy measurements. The inflation target should be zero, although yes I know that creates problems for some people–banks in particular. To heck with the velocity of money–I’ll decide when the unspent fruits of my labor need to accelerate into the stream of commerce.

  21. Gravatar of scott sumner scott sumner
    23. June 2010 at 08:49

    Mike, Reserve requirements are a tax on banks. So there is no free lunch there. You are retiring debt, but implicitly the banks are paying a tax to do it.

    Christopher. I guess they made better loans, and gradually reduced their lending during the Depression. The money supply fell as lending fell.

    pacer, Sounds simple.

  22. Gravatar of Mike Sandifer Mike Sandifer
    23. June 2010 at 09:13

    My assumption was that banks could also sell their treasuries, and reserves would only be raised to keep money velocity constant. I guess this might be too cute a view on it.

  23. Gravatar of Mike Sandifer Mike Sandifer
    23. June 2010 at 09:17

    The tax in that case might be offset.

  24. Gravatar of George Selgin George Selgin
    23. June 2010 at 15:10

    It seems to me that Doc Merlin knows more about the gold standard and competing currencies than some of his critics do. It is facile, first of all, to claim that the gold standard gave rise to more output volatility than the fiat standard has. In fact, the more careful statistical studies show very little if any reduction in output volatility since WWII as compared to before WWI. All studies agree that the interwar period was the worst of all. But to blame that on “the gold standard,” ignoring in doing so the machinations of central banks (starting with the suspensions of WWI and continuing with the house-of-cards gold exchange standard) is just plain bad science.

    Further, if you decompose the volatility into distinct supply and demand shock components, the pre-WWI period starts looking better still, for by far the greater part of volatility then was related to supply (especially agricultural) disturbances, whereas since WWII demand disturbances–you know, the ones that the fed is supposed to combat–have ruled thew roost. (I have the stats for anyone who wants them.) And this is for the US, which had a really crappy banking system (no fault of the gold standard, that) throughout the 19th century. Merlin quite rightly indicates that Canada had much smoother sailing then, as well as in the 30s (and since!) thanks to its superior banking arrangements.

    Finally, the notion that allowing private issuers to supply paper currency and coin must mean varying exchange rates _et hoc genus omni_ is a tiresome fairy tale. Consider Canada again: until it turned to central banking in ’35 (as a sop to Western social-credit types) much of its currency (save the Dominion notes) was supplied by a score of independent private banks. Yet by the 1890s it all traded at par from coast to coast. The same was true in Scotland and other places with private notes. In the U.S. branch banking barriers prevented this from occurring. Yet even there, on the eve of the Civil War, had someone bought the whole stock of state bank notes at face value and then sold them for their market value in either New York or Chicago, the lose would have been <1%.

    So, if you want to talk about facts and myths, first get your facts right.

  25. Gravatar of Mike Sandifer Mike Sandifer
    24. June 2010 at 08:25

    Scott, you replied: “Mike, Reserve requirements are a tax on banks. So there is no free lunch there. You are retiring debt, but implicitly the banks are paying a tax to do it.”

    Would it make a difference overall if the Fed paid sufficient interest on reserves?

  26. Gravatar of ssumner ssumner
    24. June 2010 at 09:37

    George, Yes, I agree that the gold standard was abused during the interwar period. My view is that either gold or fiat standards can work pretty well or pretty poorly, depending on whether governments are responsible. Merely pegging the price of gold doesn’t stabilize the price level, nor does moving to fiat money prevent gross errors in monetary policy.

    I think it is easier to run sound monetary policy with a fiat system than a gold standard, (as you don’t have the gold price peg constraint.) But most of the problems under both regimes reflect not the underlying nature of the regimes, but how they were abused by governments.

    Mike, Yes, you could pay interest on reserves. Then it would no longer be a tax. But your gain from extinguishing national debt would be exactly offset by the loss from the interest you paid. Reserves would essentially be a part of the national debt.

  27. Gravatar of Mike Sandifer Mike Sandifer
    25. June 2010 at 23:08

    It is not my intention to drag this out, but does it really matter if the reserves grow if they never make it into the economy? Might there be a way to at least slow the rate at which these reserves can be released into the economy, to minimize the costs of retiring federal debt? Would there be no way to cancel some of these reserves without discouraging banks(and perhaps the general economy) from keeping appropriate bank reserve levels in the future?

  28. Gravatar of ssumner ssumner
    26. June 2010 at 06:33

    Mike, Reserves are cash, a net asset for banks. You can’t just “cancel” them. If you mean reduce the stock of reserves by buying them back–then the answer is yes. But that is costly.

    I’m still not sure where you are going with this, all I can say is reserves don’t offer any end run around the debt problem.

  29. Gravatar of Mike Sandifer Mike Sandifer
    26. June 2010 at 14:52


    I am perhaps likely being stupid, but I have a sense that what I’m suggesting is not the equivalent of a perpetual motion machine. That is, it’s not really a free lunch, though the government borrowing may be free. There are still limits to growth rates, etc. It seems the problem needs to be looked at differently. It’s useless to express more specifically until I have some firmer ideas.

    More generally, I have this sense that some very fundamental macroeconomic ideas are flawed in deep ways. This is worth nothing to anyone else, and while I obviously have no expertise in the subject, this continues to gnaw at me. The only advantage I have is an increasing lack of rigid thinking concerning the fundamental principles.

    It wasn’t my intention to waste your time.

  30. Gravatar of ssumner ssumner
    27. June 2010 at 07:04

    Mike, No worries, your comment was perfectly reasonable.

  31. Gravatar of Sintetia » Lecturas recomendadas Sintetia » Lecturas recomendadas
    8. May 2012 at 14:59

    [...] ¿Por qué fallan a menudo los bancos centrales? Según Scott Sumner, dos mulas (la política fiscal y la monetaria) tiran a menudo en dirección contraria. [...]

Leave a Reply