The myth of the omnipotent central bank

"Inflation", said Milton Friedman, "is always and everywhere a monetary phenomenon". Upon this statement has been built three decades of faith in the omnipotence of central banks. It does not matter what government does: it does not matter what markets do: it does not matter what shocks there are to the economy. As long as central banks get the money supply right, there will be no inflation. Or deflation. Growth will simply proceed smoothly along a pre-determined path.

Leaving aside the question of whether central banks really control the money supply at all in an endogenous fiat money system, it is clear to me that the control of inflation - in all its forms - is by no means so simple. Despite Friedman's statement, the forces that create inflation and deflation are in reality poorly understood. The delicate balance between supply and demand in a monetary economy is easily disrupted: shocks to supply or demand reverberate around the economy in a manner similar to a tsunami, overwhelming everything in their path and causing lasting devastation. The means by which shocks propagate themselves are a matter of considerable debate in economic circles. There is no agreement whatsoever on what the transmission mechanisms are, let alone how to manage them. And there can therefore be no basis whatsoever for the blind faith that central banks can at all times control inflation. Or growth, for that matter (I'm looking at you, NGDP-targeters).

To be sure, central banks can limit the immediate effects of shocks. I am no fan of QE, but even I have to admit that its use by central banks around the world to prevent catastrophic debt deflation after the fall of Lehman was successful. But I would regard this as firefighting. Central banks were able to put out the bush fire caused by the fall of Lehman. But they were then expected to restore the forest to its previous green and growing state using exactly the same tools that they used to put out the fire. Flooding the place douses the flames, but it doesn't help new life to grow. On the contrary, it may actually drown it. Too much liquidity is as dangerous as too little.

What was needed after the Lehman bush fire was careful nurturing of a very damaged private sector by governments. And to start with, that is what happened. The ground healed, the plants started to grow again, and the animals began to return.

But then governments got scared. They didn't do the serious pruning of overgrown banks that was really needed to reduce the likelihood of another bush fire: instead, they encouraged central banks to flood the fire-damaged banks with liquidity in the hopes of getting them to flower again before they were really healed. But even worse, they decided that pouring water and fertilizer on the areas where fire had scorched the earth and life was slow to return was too expensive. So they stopped nurturing the scorched areas, and instead imposed a drought: the green shoots died and the animals left. They blamed the fire-damaged banks for this: if only they would flower, the plants would grow in the devastated areas even though there was no water, and the animals would return even though there was nothing for them to eat.....

And it worked, sort of. The fire-damaged banks did flower again, eventually. But the flowers they produced were not the life-giving variety that give nectar to the bees and food to the animals. They were toxic. All they fed was asset prices.

But the UK government thought this was wonderful. If asset prices rise, then that's new life in the economy, isn't it? Never mind that millions of people are unemployed or under-employed, real wages are stagnant and productivity is on the floor. The stock market is at an all-time high and London house prices are rising at over 10% a year. House prices elsewhere are rising too, because in addition to the liquidity with which the central bank flooded the banks, the starry-eyed government is adding fertilizer to the housing market.

There are now growing calls for the central bank to intervene to calm down the housing market. Central banks that have macroprudential responsibility, as the Bank of England does, have a wide range of tools that they can use to influence specific sectors: monetary policy is a blunt instrument that should be used only as a last resort. But that's not what the hawks think. A sledgehammer is the only tool they know. Therefore, because London house prices are reaching for the skies, interest rates must rise.

This is perverse. Firstly, as we have seen with the Eurozone, when monetary policy in a currency union is dictated by the needs of one particular area, the rest of the union suffers. The ECB has been under pressure to raise interest rates because of house price inflation in Germany: to its credit, so far it has resisted this. It would be a tragedy if the Bank of England gave in to equivalent pressure to raise interest rates because of house price inflation in London, and squashed the UK's fragile recovery.

Secondly, it is by no means clear that raising UK interest rates would have any effect at all on the London housing market. The principal drivers of the London housing bubble are overseas cash buyers, many of them financed at very low interest rates by foreign banks, and UK investors diversifying into property in the search for yield. A rise of a few basis points in interest rates is not going to be enough to make the returns on financial assets more attractive than London property. To have any significant effect, an interest rate rise would have to be several percentage points - which would be an enormous negative demand shock to the UK economy. Using monetary policy to prick the London housing bubble would be likely to knock the economy back into severe recession.

And thirdly, it is not only perverse but morally wrong for an unelected central bank to attempt to override or counteract the intended economic effects of the policies of an elected government. Help to Buy is pushing up house prices outside London: this hurts the people it is primarily designed to help, but it does encourage construction, and the rising house prices create "wealth effects" which act as a much-needed economic stimulus in depressed areas such as the North East. Despite widespread derision at what appears to be blatant buying of homeowner votes, the government has made it clear it has no intention of ending Help to Buy. Therefore we must assume that the government likes the current housing boom. While Help to Buy remains in place, the Bank of England has no business raising rates purely to calm the housing market. If financial stability is threatened, Help to Buy must end.

I doubt if macroprudential tools would have much effect on the London housing market, since it is primarily driven by cash buyers. Fiscal tools would be far more effective, and there are numerous alternatives. Punitive taxation of capital gains on speculative purchases: 100% taxation of notional rents on properties left empty: wealth taxation on very high value properties: a top band on council tax: even a land value tax. Anything to give the clear message that housing is first and foremost shelter, and speculative activity that drives up prices beyond the reach of ordinary people is not acceptable. But this, of course, is the job of government - not the central bank.

And this brings me to the real point of this piece. The myth of central bank omnipotence has been carefully fostered by the academic economics profession, by central banks themselves and by politicians. The first two of these clearly have a vested interest in technocratic management of the economy. But it is perhaps less clear why politicians sell this myth.

Politicians sell the myth of central bank omnipotence because it allows them to pursue all manner of stupid fiscal and quasi-monetary policies while escaping responsibility for the consequences. If Help to Buy causes asset price inflation, it's the central bank's responsibility to sort it out - even at the cost of widespread collateral damage. If a front-loaded austerity programme in the teeth of oil price shocks and a sovereign debt crisis in our largest trade partners squashes a fragile recovery, it's the central bank's responsibility to get growth going again. If ill-considered and callous changes to the benefits system force a lot of people on to the unemployment register when the jobs market is already depressed, it's the central bank's responsibility to stimulate the economy so new jobs magically appear. If the central bank fails to achieve these because of an unhelpful fiscal stance, blame the central bank.

But the truth is that central banks are not omnipotent, and the tools they use are not cost-free. As already noted, interest rate changes can have unpleasant and unintended side effects. The effectiveness of QE when the fiscal stance is contractionary is uncertain, and it is unquestionably regressive, benefiting the very rich far more than any other section of society. Macroprudential tools are of questionable effectiveness too: they are easily evaded or watered down, and they are ineffective if improperly targeted. And the inadequacy of information available to central banks - and poor timing - means that their actions are always likely to be "too little, too late". For targeted sector interventions, fiscal tools can be both more powerful and more effective: we should be far more willing to use them.

Most importantly of all, central banks simply do not have the power to override a determined government. One of my favourite academic papers, Sargent & Wallace's "Some Unpleasant Monetarist Arithmetic" reminds us that central banks are powerless to control consumer price inflation generated by reckless government borrowing and spending. The same is true of asset price inflation generated by irresponsible and meretricious government support of housing markets. And in my view the same is also true of deflation caused by inappropriate fiscal contraction.

Whether they like it or not, politicians are accountable to their electorate for managing the economy. It is wrong of them to attempt to pass this hot potato to unelected central banks, and equally wrong of central banks to think they can accept it.

Related reading:
Why house prices are a problem for governments, not central banks - Pieria
The temples of the gods of capital
FLS and the Bank of England's independence
Making the desert of plenty bloom - Pieria


  1. Frances, this is what NGDPLT does....

    It IS a CB becoming political, or it is CB removing itself, however you start the frame.

    But don't fool yourself, NGDPLT will force the fiscal side to focus on making sure ALL its spending = RGDP, not inflation.

    A public sector pay raise that doesn't include productivity gains will always be inflation, not RGDP.

    We are a tricky lot.

    1. I know what NGDPLT does, thanks. You've missed the point - which is one I've made before - namely, that CBs do not have as much power or as much information as you think they do.

    2. Frances, Friedman said he just wanted to replace the CB with a computer.

      If you are saying a computer program can't make NGDPLT hit 4% month to month for decades and decades, explain why.

      If you are saying, you don't want it to do so, make it clear.

  2. Hitting NGDPLT is remarkably easy.

    If we're at 100, one year later we have to be at 104.

    To keep it really simple:

    1.0033 -1
    1.0066 -2
    1.0099 -3
    1.0132 -4
    1.0165 -5
    1.0198 -6

    So if we come in at 1.0040 in month 1, the computer now takes aim at only adding .26 where the expectation was adding .33 by end of 2. we tighten.

    If we come in at 1.0017 at end of month 1, the computer now takes aim at adding .049 - we expand.

    Now if the markets FIGHT the computer

    And the next month we come in at just 1.0027, the computer is now going to go ape, adding .0072 for next month.

    The relentless Level Target, the cold brutal computer that pays NO attention tot he cries of political interests...

    that's the truly POWERLESS Fed, no?

    Whose policies are bad? Whats causing the market to fight and lore to monetary?

    OH, its this mortgage lending policy.... you get the idea.

    1. Morgan,

      I really don't need you to explain to me how computer systems work. I was designing computer systems when you were in diapers. But you ignore the human dimension. Economic shocks are created by humans, who behave irrationally. And sometimes they are created by nature, which behaves chaotically.

      So yes, I am saying that a computer cannot make NGDP hit any target. It can adjust policy in response to lagging indicators that tell it what was actually achieved and how far away from target it was, and it can attempt to forecast what NGDP will be using the best information it has available at the time. But what NGDP actually is in reality may or may not bear much relationship to the forecasts.

      This does not mean that rule-based monetary policy is useless. Far from it - it's a useful framework. But you should be aware of its limitations.

  3. You are missing the point.

    Take NGDPLT at 4% out 10,20,30,50 years.

    We all know the EXACT target for every month, every year.

    Sumner uses the metaphor of a boat captain. Think of the computer like this.

    The island is right there 20 years away, you can see it. Each month you will move the wheel directly towards the island. Sure you might be a little over a little under, but since you are ALWAYS going back to the level target.

    There is no guessing by anybody what the GDP will be in 20 years.

    Now the ONLY question is whether the economy will be great - lots of real growth, or terrible with lots of inflation.

    And that means we will blame the politicians if it is mostly inflation, the guys you want to take responsibility! And if fiscal instead delivers mostly real growth, the politicians will get the credit.

    Or get out graph paper and place a dot, and 120 spaces away place a random dot, not without a ruler, draw a line on box and stare at it. Then do it again.

    You WILL "in the long run" hit the dot, you will hit the island.

    But Frances, SO FAR to date, there hasn't been a level target.

    If we had a 2% inflation level target since 2008, then the computer would have long ago increased inflation to get back to the level target.

    So the system you are talking about, hasn't yet been level. The system you are saying can't work, hasn't been tried.

    And I know this for sure, many politicians will NOT LIKE NGDPLT, and you don't like it either, bc they will not be able to pursue the kind of not productivity gain government spending you favor.

    MM turns government that is less productive than the private sector into "inflation" - and you hate it.

    1. Morgan, you are still entirely missing the point. The economy is a human system and humans are irrational. And the economy is founded in nature, which is chaotic. Unless you can find ways of modelling irrationality and chaos so accurately that NGDP forecasts are spot-on, you will never "hit" an NGDP target. You may come close to it - or you may not.

      What I personally "feel" about NGDP targeting is irrelevant. And you accuse me of wanting unproductive government spending, for which I have NEVER argued. Instead of making wild accusations, you should actually bother to read what I have written.

  4. You are having feelings.

    I get what you have written, and I am saying you are wrong.

    Right now the fed has a "don't go over 2%" approach, and they do a pretty good job of it. because even having just 1% meets their rule.

    If they had a 2% level target, and they did the same quality of job at that, we'd have for short periods a higher target like 2.5% or a lower target like 1.5% annualized for any given month.

    So if in month 1 you come in low at 1/12 of 2%, your month 2 target is 1/12 of 2.5% because you have to MAKE UP the miss.

    If in month 3, you come in high over 3/12 of 2%, then you try to add only 1/12 of 1.5%.

    So you are buying selling assets increasing decreasing interest rates to keep getting back to level.

    The ISSUE is you aren't saying WHY the new policy won't beat the status quo, you are just saying "humans are fallible"

    And ok sure yes (we want to use a computer) but they are ALREADY fallible in the status quo, the question you not answering, the analysis you are not bringing is:

    Why won't the NGDPLT beat the current system, assuming the same fallible humans in each?

    1. I have not suggested that NGDPLT is not better than inflation targeting.

      I have said it is inevitably less than perfect, because humans are irrational and nature is chaotic.

      Why do you have a problem with that?

  5. I think we're getting somewhere. Which would be nice. :)

    1. I "think" you'd agree level targeting whether ngdp lt or inflation lt is better than not having a level target ("no make up" like we do now), but I'm not sure, so thats my first question.

    2. I'm not sure if you'd prefer ngdplt versus inflation level targeting.

    3. This is the big one, do you see what MM mean when we say a level target removes human decision making from the equation? Its a step closer to Friedman's computer. Do you disagree for some reason with our belief here?

    4. If you want to put the responsibility back on politicians to create good fiscal policy, and I want to see this happen to, I can't figure out why you'd not want to remove discretion from the Fed...

    1. Morgan,

      Actually, we aren't getting anywhere at all. Your comments are WAAY off topic and you persistently ignore my attempts to steer you back to the subject. This post is not about the relative merits of inflation versus NGDPLT targeting. It is about the limits of central bank power. Please confine yourself to the topic of this post.

    2. Let me repeat back what you are saying, so you know I understand it:

      You start with Inflation isn't just a monetary phenomenon. I let this go, because while its not right, I don't get much from having you admit it.

      You've actually made my and Milton's argument yourself - you wrote about Bitcoin, you mentioned there is no inflation in it, you even USED this to question its moneyness.

      Milton doesn't mean anything other than you said yourself.

      2. You are correct that there are non-monetary shocks that can affect inflation / deflation, but Milton wasn't ignorant of this, and certainly you and I agree at length about the deflation forces of digital.

      So on #2, coming thru your second paragraph - it's still a more basic confusion. Milton didn't mean the Central Bank was omnipotent, he meant you don't get inflation unless you are growing the money supply. Many say they equal each other.

      Again, I repeat, milton did not think the CB was omnipotent, he thought they were fallible idiots beholden to political interests.

      think of monetarists generally as Austrians, who are pragmatic enough to say "OK, lets grow money" so we can keep the economy going like Frances agrees - so we don't have to put up with government doing fiscal - bc we'll reduce the number of crises.

      i REALIZE that you want to have the political class manage both fiscal and monetary. But thats not interesting.

      3. You start talking about housing markets and again I basically skip it, because London has terrible land use policies, and that's not a MONETARY problem - see what i did there????

      What i did was take monetary off the table, to force you to own the real fiscal problem - bad land laws.

      By removing your ability to mess with money supply, you are left with MASSIVE INFLATION in housing in UK, which is unhealthy, and forces you to confront your REAL PROBLEMS.

      4. Then you explain rightly, that first human discretion at the CB allowed home prices to rise,and then you note there are calls to STOp monetary to cool off housing.

      This is again very interesting, but ONLY because it indicates that removing human discretion will keep there from being calls for the CB to do ANYTHING.

      That's where I came in.... The organic problem if BAD FISCAL.

      letting ANYBODY scream at the CB to react to bad fiscal, just bollocks it up.

    3. I repeat in 1-4, you are wrong about Milton saying CB is Omnipotent. What he WANTED to do is remove discretion at the Fed. You are not confronting his true goal. He wanted to remove discretion, so that the London housing boom / bust could not be driven by calls for CB discretion. he wanted a cold computer to force your government to change its land use policies, whether they like it or not.

      Now Frances, you might not think it is RIGHT for monetary policy to no longer bend to the will of government policy, but NOWHERE in this piece do you make the argument, or confront directly what the real Monetarist and NGDPLT goal is.

      In many of your last paragraphs you again attack straw men - NOBODY on my monetary side wants CB discretion. You list countless problems with Monetary that come from having discretion.

      I'm not going to OWN THOSE - I do not OWN THOSE - you OWN THOSE, because you do not seem to support removing discretion from the CB.

      Anything that comes from CB discretion that you complain about - is not an attack on NGDPLT or Milton Friedman.

      Now please stop saying I don't read and understand you. You realize that is nuts right? I'm a pretty good debater, I've been coaching it on and off for years and years, I'm able to flow computer logic and architect complex software, and we do not disagree because I don't get what you are saying, we disagree bc you are hiding some giant assumptions in there.

      That humans beings are fallible is not an argument against free markets, NGDPLT, or anything else. it is an argument against strong central authority, it is an argument for anti-fragile systems.

      That your CB has discretion is not an argument against CB, it is an argument against discretion.

      And none of it gets down to this example: you have band land policies, they are causing problems, and NGDPLT is meant to force you to fix them, and you don't want to not get to use MP AND FISCAL to try and keep the govt policies you like but the markets don't.

      You think society should set the laws the way it wants and use govt spending AND money policy in concert to make it work.

      And that is a terrible idea.

    4. Morgan, what is it about my clear statement that this post is not about the merits of NGDPLT targeting that you don't get? And what is all this stuff about Friedman? I quoted Friedman because I happen to think his explanation of inflation is simplistic - as is yours, by the way. But nowhere did I say that Friedman said central banks were "omnipotent"?

      Actually this post is a thinly disguised attack on the UK's Chancellor of the Exchequer for his awful fiscal policies. But you have completely missed that, haven't you? And that's because you ignored half my post because you wanted to set up and knock down a straw man of your own. You seem to think that I WANT awful fiscal policies. No I don't. I am objecting to the assumption - on the part of the UK's Chancellor, among others - that central banks can always compensate for fiscal disasters. If you had bothered to read Some Unpleasant Monetarist Arithmetic, you would know that they can't. I do put links in posts for a reason, you know - they are there for you to read to help you understand what the post is about.

      You seem to think that I want fiscal to dominate monetary. I certainly don't - I made that crystal clear. But that doesn't mean there is no role at all for fiscal policy. The imbalances in the UK's housing market would respond to targeted taxation far better than to monetary policy. In this post am countering growing calls for the Bank of England to use monetary policy to calm a specific housing bubble in London caused by low rates and inflows of hot money. In my view this would be a completely inappropriate use of monetary policy. With your commitment to NGDPLT targeting, I would have thought you would agree with me.

      So no, you don't read me properly and you don't understand me. In fact you have utterly misunderstood this post. Before you comment any more, I suggest you actually take the time and trouble to read this post PROPERLY, including all the links. And stop assuming this is an attack on your ideology. It isn't anything to do with it.

  6. See AGAIN, we're so damn close!

    When you say "that central banks can always compensate for fiscal disasters."

    You misunderstand MM - we do say our god can lift any rock. But we don't say that bc our god SHOULD lift any rock.

    Instead, we say

    1) Money Policy can always hit the target if it wants to - our god can can lift any rock.

    2) BUT we only say that to keep big govt. Keynesians the pretending from saying we can't.

    3) THEN because we hate MP being used by politicians for anything, We then say - but we'll use a level target, say 4% - and our computer / futures market god will hit that thing rain or shine like John Bonham:

    4) SINCE we have taken MP out of government hands, they will have to fix housing policy. They will have no choice.

    The issue is that I want to TRULY force them to fix Fiscal, and you want them to keep control of CB, but use fiscal instead, and stop blaming CB.

    You know when we run tests, we want to control the variables, NGDPLT reduces all problems to Fiscal problems - it does so much more aggressively than you are calling for.

    Thats why I keeps asking about level targeting vs. NOT level targeting - because thats a way I think you'll agree removes discretion...

    When your politicians try to blame MP, they will be blaming 4% NGDPLT.

    And the answer back from the computer is:

    If you are creating inflation instead of real growth in the 4% - that's your fault, and I'm a computer / futures market.

    and whichever politician creates policies that make it 3% RGD and 1% inflation he'll get reelected, and the ones that make it 3% inflation and 1% RDGPO, they get fired.

    And suddenly we all have to admit which policies create inflation and which ones create real growth.

    NGDPLT is a debate ender. It forces us all to run tests and get evidence, because the metronome of our music never misses.

    1. Morgan,

      I think your refusal to engage with the issues in this post is wilful. This conversation is therefore at an end. I will not post any more comments from you.

  7. You don't like John Bonham?

    No really, I think we're there almost.

    I don't think you want "Fiscal to dominate monetary", I think you want government to have control of monetary, maybe not blame it, but not give up control.

    I'm saying govt. should have no control over monetary and I have no idea why you won't agree to that.

    1. No, you have COMPLETELY missed the point. Totally and completely.

      There is no point in any further discussion.

    2. you know you are stuck...

      I'm running a counterplan that says take discretion of MP away from govt / CB. and you have gone long on saying how bad CB is...

      And you don't really want govt. to lose control of MP discretion. And you won't say why.

    3. Crikey. How have you managed to interpret my post as meaning exactly the opposite of what I actually said?

      As I said, there is no point in further discussion. This conversation is at an end. Further comments from you will not be posted.

    4. Ms. Coppola, first I just want to say how much I love your movies almost as much as this blog. Second, that conversation looked like a wackamole merry-go-round from hell. You made the right choice.

    5. Haha. I'm relieved that you like my blog more than my movies....

  8. I think that you understate it when you say "the forces that create inflation and deflation are in reality poorly understood" - in fact for inflation it's clear that we don't even have a particularly good definition of what it is. (I suspect the same is true for deflation, but there have been less episodes to examine.)

    The shortest way to put it is, if (as is often said) we don't have a good way to identify a bubble, then we also don't have a good way to identify inflation. This shows up most obviously once we get away from the classic wage-price spiral. In the last few years in the UK we've seen inflation related to VAT rises, fuel rises etc. - and no wage rises. Yet most discussion of inflation mechanisms run through consumer spending and wages...

    1. Yes, perhaps I should simply have said "not understood". For forms of inflation other than consumer price inflation this is probably close to the truth. And we REALLY don't understand deflation. Or hyperinflation. Consequently we have quasi-religious beliefs around both of these that are persistent and irrational, founded on inaccurately remembered and misunderstood memories of past bad experiences. We desperately need to slaughter some sacred cows.

  9. Frances,

    As ever a wonderful post and thread - not too sure what our Monetarist friend was babbling on about, but there again I no economist thank God.

    That said, I'm in London working on a gathering in November that should be of great interest too you - its on banking supervision and of a senior level, how then does one communicate with you to make sure I can make you a guest of mine and get you to contribute from the audience?

    Christopher D. Rogers

    1. Hi Christopher. Ping me an email on and we can discuss.

  10. Frances,

    Many thanks, details have been forwarded and looking forward to greeting you and a few other commentators at the gathering in November.


Post a Comment

Popular posts from this blog

WASPI Campaign's legal action is morally wrong


What really happened to Signature Bank NY?